Talking to Skeptics

Home , Aggressive mood, Essays (Montaigne)

TALKING TO SKEPTICS

by Charles Goldhaber B.A. in Philosophy and Japanese Language & Culture, summa cum laude University of California, Berkeley, 2011

Submitted to the Graduate Faculty of the Kenneth P. Dietrich School of Arts and Sciences in partial fulﬁllment of the requirements for the degree of Doctor of Philosophy

University of Pittsburgh 2020 UNIVERSITY OF PITTSBURGH

DIETRICH SCHOOL OF ARTS AND SCIENCES

This dissertation was presented

Charles Goldhaber

It was defended on

October 29, 2020

and approved by

John McDowell, PhD, Distinguished University Professor, Philosophy

Stephen Engstrom, PhD, Professor, Philosophy

James Shaw, PhD, Associate Professor, Philosophy

Karl Schafer, PhD, Professor, Philosophy

Dissertation Director: John McDowell, PhD, Distinguished University Professor,

Philosophy

ii Copyright c by Charles Goldhaber 2020

iii TALKING TO SKEPTICS

Charles Goldhaber, PhD

University of Pittsburgh, 2020

Skeptics argue that we can know very little, threatening our claims to knowledge. Many philosophers now think there is no use in talking to skeptics—that nothing can change their minds, curing them of skepticism. These philosophers opt for mere prevention, aiming to convince only non-skeptics that skeptical arguments fail. I argue that a cure is needed, viable, and theoretically illuminating. First, I argue that the mere prevention of skepticism is likely to fail. To succeed, it would need to show why skeptical arguments appear compelling. Showing this either reveals that the arguments are compelling, demonstrating the need for cure, or reveals how they merely appear so, constituting a cure. Second, I show that we can change a skeptic’s mind. I argue that influential arguments for skepticism about the external world all rely on a shared, tacit premise: that perception never guarantees that things are as we seem to perceive them to be. I then argue that arguments for this premise are question-begging. Showing the skeptic that her skepticism lacks foundation clears obstacles to her accepting, on ordinary grounds, a view on which perception can provide us with knowledge of how things are around us, thus curing her. Third, I argue that my cure helps us understand the nature, significance, and history of skepticism. For Hume, I explain, skepticism is a temperament, which leads to madness when overly dominant, but only carefulness when balanced with other temperaments. Tracing skepticism to a groundless intuition helps motivate Hume’s focus on temperaments, while Hume’s conception of proper temperamental balance helps to diagnose and moderate skepticism. I then argue that Kant tries to cure a skeptical empiricist not by showing skepticism’s incoherence, but by offering an alternative explanation of our knowledge. Kant’s portrayal of skepticism as arising from despair of understanding human knowledge explains why the skeptic is apt to find his alternative appealing, and teaches a general lesson about curing

iv skepticism: Oﬀering the skeptic a way to make sense of our knowledge allows him to overcome the frustration from which his skepticism arises.

Keywords: skepticism, knowledge, perception, temperament, Hume, Kant.

v TABLE OF CONTENTS

PREFACE ...... ix 1.0 CURE AND PREVENTION ...... 1 1.1 WHO SKEPTICS ARE...... 1 1.2 KINDS OF RESPONSES TO SKEPTICISM...... 3 1.3 THE PREFERENCE FOR PREVENTION...... 8 1.4 IGNORING THE SKEPTIC...... 11 1.5 A DILEMMA FOR MERE PREVENTION...... 16 1.6 TRANSITION: HOPE FOR A CURE...... 21 2.0 THE GROUNDLESSNESS OF SKEPTICISM ...... 24 2.1 PLAN FOR A CURE...... 24 2.2 A COMMON PREMISE IN SKEPTICAL ARGUMENTS...... 25 2.3 JUST AN INTUITION...... 36 2.4 REACHING THE SKEPTIC...... 40 2.5 OBJECTIONS AND REPLIES...... 43 2.6 CONCLUSION...... 46 2.7 TRANSITION: HISTORICAL PRECEDENT...... 52 3.0 THE HUMORS IN HUME’S SKEPTICISM ...... 56 3.1 AN OVERLOOKED QUESTION...... 56 3.2 A CYCLE THROUGH THE HUMORS...... 60 3.3 THE ANACHRONISM OBJECTION...... 69 3.4 THE METHODOLOGY OBJECTION...... 75 3.5 RECONCEIVING HUME’S SKEPTICISM...... 80 3.6 CONCLUSION...... 88 3.7 TRANSITION: JUST A PALLIATIVE?...... 91 4.0 KANT’S OFFER TO THE SKEPTICAL EMPIRICIST ...... 96 4.1 A NEW QUESTION...... 96

vi 4.2 TRADITIONAL READINGS OF KANT ON HUME...... 101 4.3 KANT’S ATTITUDE TOWARDS SKEPTICAL EMPIRICISTS...... 111 4.4 THE EXPLANATORY READING OF THE DEDUCTION...... 116 4.5 THE EXPLANATION IN THE DEDUCTION...... 122 4.6 THE INSTABILITY OF SKEPTICAL EMPIRICISM...... 131 4.7 FURTHER OBSTACLES...... 144 4.8 CONCLUSION...... 148 5.0 CONCLUSION TO THE DISSERTATION ...... 151 APPENDIX A. CHAPTER SUMMARIES ...... 158 APPENDIX B. NOTE ON CITATIONS OF HUME AND KANT ..... 160 BIBLIOGRAPHY ...... 163

vii LIST OF FIGURES

1 The four humors...... 62 2 Abbreviations for Hume and Kant’s writings...... 162

viii PREFACE

This dissertation is a set of interconnected essays on the theme of skepticism. Common to each essay is the idea that we will understand the nature and proper treatment of skepticism better than many contemporary discussions do, if we consider the history, ambitions, moods, and feelings which give rise to skepticism. Throughout each essay, I will be treating skepticism as a kind of infirmity or illness. As an illness, its treatment is aided by looking beyond its symptoms and for its underlying causes. In viewing skepticism in this way, I do not mean to disparage or trivialize the phenomenon, those afflicted with it, or the thinking that it can inspire. On the contrary, I think we must take skepticism seriously. We must treat it with a sober respect in part because, like some diseases, it is a regular and natural part of human life, and can be brought about through human activities that are often or in other circumstances salutary. Viewing the illness in light of these origins can be one way of coming to see skepticism as capable of playing a positive role in human life, especially insofar as it can help usher in a kind of intellectual maturity. Nonetheless, skepticism’s harmful side should not be overlooked, and the production of a cure for such an illness is a worthy humanitarian effort. One of the principal goals of this dissertation is to offer a cure for skepticism. I develop this cure in Chapter2. But another goal is to diagnose what I see as a deficiency in recent philosophical discussions of skepticism, which I believe tend to distort the nature of the phenomenon by abstracting it from human life. Most philosophical discussions of skepticism nowadays treat it as one of many possible, equally well-positioned views about human knowledge or justification. In particular, they take skepticism to be the view that we know nothing, or very little, of what we ordinarily take ourselves to know, or that we are justified in believing little to none of the things we believe. Like any other view in epistemology—say, fallibilism or contextualism—arguments are produced in its favor, and considered on their own merits, presumably from a hallowed position of philosophical neutrality. This sort of approach to skepticism may appear appropriate, insofar as philosophy itself can seem to require sober non-partisanship. Surely part of what makes philosophy attractive

ix and worthwhile is that it can, in some circumstances, not only help us to understand the grounds of what we antecedently believe, but also change our minds and make us wiser as a result. To succeed with this, philosophy may indeed require us to suspend judgment about the topic of interest and patiently turn it on all sides before deciding what to think. But thinking that philosophy must always do this, or even that there can ever be a perfectly pure decision from the void, is, I think, a mistake. This way of thinking removes philosophy from its actual, humble origins in human concern and insight, and, in doing so, impoverishes the resources of philosophical dialectic. It can lead us to overlook the possibility of appealing to more than our interlocutor’s explicit commitments. We might then miss the relevance of her antecedent aspirations, affective tendencies, and ability to dissociate from her current views sufficiently to entertain others and see where they lead. It may seem that it is skepticism in particular, rather than philosophy in general, which demands that we limit the resources used in debate with a skeptic to a very narrow set, insofar as skepticism casts doubt on so much of what we ordinarily take ourselves to know. But to think that this is so, I suggest, is to forget that skepticism is a condition that people can be afflicted by, and that philosophy can and ought to play a role in the treatment and moderation of such natural infirmities. In order to reclaim a more resource-rich dialectic, I propose to model philosophical treatment of skepticism on a conversation with a skeptic. In calling my dissertation “Talking to Skeptics,” I do not mean that I will be using a dialogical form. Instead, I call it that to remind the reader that skepticism is something a person has, and so the treatment of skepticism is in part the treatment of a person. And that means that any features of that person—his aspirations, moods, and tendencies—which lead into his skepticism can properly figure within a philosophical diagnosis of the phenomenon. So, going forward, I will speak primarily of skeptics, rather than skepticism, in part to make clear that a treatment of skepticism may address and exploit the human elements which turn up in the illness’s natural history. Doing so, I hope to show, can combat an impoverished contemporary view, on which skepticism appears both more indomitable and less interesting than it actually is. Insofar as this contemporary view can make skepticism appear virtually irrefutable and simultaneously inert, I think it has given rise to a recent and growing sense in philosophy

x that there is no use in talking to skeptics. Many philosophers now believe that the most viable reaction to skepticism available to non-skeptics is to satisfy oneself with the thought that skepticism was never really a threat in the first place. I find this trend unfortunate not just because I think it is unlikely to succeed, as I suggest in the introductory Chapter1, nor even primarily because I doubt that it is a common-sense thinker’s only option, as I argue in Chapter2. I think contemporary discussions also miss out on some of the positive effects of talking to skeptics, and even of contracting and recovering from skepticism. I discuss this beneficial side of skepticism in Chapters3 and4. These two chapters look to the great modern philosophers David Hume and Immanuel Kant to draw out some of these positive results. In addition, I use Hume and Kant’s discussions to show there is a precedent for seeing skepticism as a natural part of human life, and one whose treatment involves understanding skepticism’s origins in normal human drives and feelings. In this way, these latter chapters serve to justify the approach to skepticism I introduce here and develop in Chapters1 and 2. Additionally, I hope that the discussions here, and in Chapters1 and2, justify what, from the perspective of contemporary epistemology, may seem like an odd choice of passages to focus on in my readings of Hume and Kant in Chapters3 and4—passages which often emphasize moods, feelings, and connections to medicine. In that regard, the first half of the dissertation may help to reveal why my readings of these philosophers are novel, or at least unfamiliar in our time. This preface has so far emphasized the connections between the chapters. But, in what follows, I have tried as far as possible to make each chapter a self-standing essay. Readers who are interested in only one chapter may then read it in isolation, without fear of completely lacking context. To accommodate this structure, I have included “Transition” sections at the end of most chapters, which pause to explain the relevance of one chapter to the next and to the dissertation as a whole. These may be skipped by anyone reading a chapter in isolation from the others. But even such readers may find the “Transition” sections interesting, if only as advertisements for the other chapters. Such readers may also find useful the short “Chapter Summaries,” which begin on p. 158. For all intents and purposes, these summaries are functionally equivalent to abstracts. Readers interested in comparing my readings of Hume and Kant to their primary texts will find useful the Appendix on the abbreviations

xi and editions used for my citations of these figures, beginning on p. 160. Before moving on to the body, I want to address two important points. The first concerns skepticism as a threat to our knowledge. Some philosophers view skeptics as questioning the justification of our beliefs, rather than their status as knowledge, and likewise aim only to defend justified belief rather than knowledge. But I will not be following this trend here. Skeptics, as I view them, traditionally doubt the ways in which we ordinarily take ourselves to be in touch with an objective reality. While some may attack our knowledge by arguing that our beliefs are not justified, others may stress that what, from our point of view, we are perfectly justified in believing may nonetheless be false. In either case, we can view the skeptic as posing a threat to our knowledge. And in either case, defending our knowledge is a worthy pursuit. Moreover, defending only our justified belief may reflect a concession to the skeptic: namely, a concern that, after appreciating the skeptic’s arguments, we cannot really get knowledge or an objective reality back into the picture. I do not think we need to make this concession. On the contrary, a retreat to justified belief is, on my view, a form of skepticism, or at least a symptom of its taking hold. Second, I claim to be ‘talking to skeptics’ and offering them a cure. But I imagine many readers may wish to ask at this point: Are there really any skeptics among us of the sort discussed in philosophy? Does anyone really believe that she knows nothing, or doubts that she knows much of anything, including even the most basic things about the world around her? Do we not treat a great many things as known as soon as we, say, go to make breakfast or search for our house keys? Do we not immediately take ourselves to know that the eggs are in the refrigerator or that our house keys hang from a small, metal carabiner? Even if it is correct to say in those instances that we do believe that we know a number of things—which strikes me as a very plausible thing to say—I still think there can be, and in fact are, many skeptics of a very general sort. Such people may perhaps be said to only partially or intermittently hold that we can know nothing. But they can still hold that view stably enough to count as skeptics. They may, for example, have a propensity to report that we can know nothing, when asked about their views on human knowledge. Again, such people may be impressed by arguments for skepticism, and as yet see no way to avoid drawing their conclusion. They may carry with them a feeling of discomfort, which, though perhaps

xii often unpronounced, easily emerges when they are asked to consider what they know, or on what grounds they believe the things they do. They may, at least on occasion, doubt that the glass of water or person before their eyes is real. They may be accosted by intense suspicion at inappropriate times and feel a sense of alienation, or unease, or awe in its wake. All of these phenomena are a part of the condition I would like to call skepticism, and may be seen as at least partial grounds for attributing the view that we can know very little to any person who suffers from a cluster of them. On this conception, I think skepticism of varying degrees is in fact widespread. Many of my undergraduates suffer from it, at least for a while, after reading Descartes’ first Med- itation. Many epistemologists contract it, when they find the considerations for a skeptical view of human knowledge hard to resist or determine that there can be no cure for skep- ticsm. And, more generally, reflective and unreflective people alike nurture it in daydreams or as a result of some belabored thinking or feeling of disaffection. If I am right, even the anxiety and vertigo of modern culture can be a way into various moods that, even without argument, lead a person to groundlessly adopt a skeptical view. Viewing skepticism in this light helps to reveal the extent to which seeking a cure is a compassionate pursuit. A cure could serve to rescue some of us from a kind of malaise, and reconnect those people with the crisp reality of a world we share with others. That humanitarian concern is one reason for talking to skeptics.1 To whatever extent this dissertation expresses my own thoughts, it doubly reflects the benefit of countless conversations, in which its raw materials first took shape. I owe a great debt to the members of my doctoral committee: John McDowell, Stephen Engstrom, James Shaw, and Karl Schafer. Each has been enormously generous with his time, encouragement, and trust; has offered me consistent guidance and incisive criticism from the project’s early explorations through its final revisions; and has deepened my interest in philosophy and its history. At the risk of oversimplifying the many and various ways each advisor has influenced

1Here, I am opposing myself to, for example, Rorty(1980), 728, who views the “sceptic” as “one of the Platonist or Kantian philosopher’s imaginary playmates, inhabiting the same realm of fantasy as the solipsist. . . and the moral nihilist. . . These positions are adopted to make philosophical points—that is, moves in a game played with fictitious opponents, rather than fellow participants in a common project.” Skeptics, as I am conceiving them, are neither fictitious nor opponents; they are fellow people, afflicted with a familiar condition. Though the condition and its treatment are of interest to philosophy, their import is by no means limited to “moves” in a purely theoretical enterprise, let alone in idle games.

xiii my thought and improved my writing, I want to briefly thank each for something unique he brought to the process: Thanks to John, my committee chair, for his incredible mentorship, his attention simultaneously to the smallest of details and to the largest arcs of thought, and for his empowering confidence in the reality and familiarity of human knowledge; to Steve for his many generously lengthy conversations which helped me to see, through both their content and tone, the humane and interpersonal dimensions of Kant’s philosophy; to James for working with me to get this project off the ground and find a foothold in the contemporary literature, and for offering me the language for articulating the ways in which my thinking is novel; and to Karl for shepherding me through Hume and Kant’s systems, never shying away from their complexity, and always bringing to light their interest for contemporary thought. It has been fun. I would like to give a special thanks to Colin Chamberlain, who, though not officially on my committee, essentially acted the part. Colin saw me through more drafts of Chap- ters2 and3 than I can remember, and always with infectious chuckles of excitement. His companionship helped me cope with the growing pains of professionalization and the job market. For help with professionalization, I would also like to thank Lee-Ann Chae and Erica Schumener. Both have helped me figure out how to present my work concisely to wide audiences, and in the process managed to move my research forward. For that I am grateful. Each of the chapters has benefitted from written and in-person conversations with many other people, more than I can remember or list here. For comments on Chapter2, I thank Eli Alshanetsky, Santiago Echeverri, Joshua Eisenthal, Elena Garadja, Ned Howells-Whitaker, Adam Marushak, Aaron Salomon, Pablo Zendejas Medina, and—especially—Stephen Mack- ereth. For comments on Chapter3, I thank Donald Ainslie, Bowen Chan, Jonathan Cottrell, Kathleen Creel, Sophie Dandelet, Dominic Dimech, Vivian Feldblyum, Tyler Haddow, Jen- nifer Maruˇsić,Lisa Miracchi, Rhetta Morgan, Hsueh Qu, and several anonymous editors and referees at Ergo: A Open Access Journal of Philosophy, in which this chapter is forthcoming.2 And for comments on Chapter4, I thank Garrett Bredeson, Alessandra Buccella, Andrew Chignell, Tyke Nunez, Gregg Osborne, Jelscha Schmid, Daniel Sharp, Peter Thielke,

2As Goldhaber(forthcoming-a). I retain copyright under the Creative Commons Attribution- NonCommercial-NoDerivs 3.0 license, so no special permissions are needed to include the material here. But Ergo deserves a special thanks for their exemplary open access model.

xiv and Alnica Visser. I would also like to thank Sophia Arbeiter, Josh Fry, Anil Gupta, Jim Hutchinson, Daniel Kaplan, Japa Pallikkathayil, T. Raja Rosenhagen, Jack Samuel, and Robert Steel for their comments on work that helped me make my way to this dissertation, but does not appear within it. Many of the folks I have just listed are or were part of the University of Pittsburgh’s amazing Philosophy graduate student community. It has been a great privilege to nurture my thoughts, and my soul, amongst such peers. I am especially grateful to Will Conner, Haixin Dang, Joshua Eisenthal, Vivian Feldblyum, Elena Garadja, Lok-Chun K. Gustin Law, S.M. Love, Stephen Mackereth, Aaron Segal, Maximilian Tegtmeyer, Gabriel Vasquez- Peterson, and Alnica Visser. Through many conversations over dinner or cups of tea, each has shaped my philosophical identity, and, more importantly, my person. They brought companionship and laughter to my years in Pittsburgh. I also want to acknowledge those I see as leading me onto the path of philosophy. In- strumental in this were my late undergraduate mentors at UC Berkeley, Hubert Dreyfus and Barry Stroud. Bert’s courses on existentialism, literature, and Heidegger encouraged me to leap into an unconditional commitment to the discipline. From Barry’s teaching and writing I grew an appreciation for the topics, figures, method, and style that I have adopted in this dissertation. Both philosophers held in view, in very different ways, the thought that philosophy is a human activity and is answerable to the phenomena of lived human experience. For me, they brought philosophy down to earth. I am also grateful to Hannah Ginsborg and Niko Kolodny for teaching me how to read, write, and think. And, for similar reasons, I am equally grateful to my undergraduate peers Laura Davis, Andrew MacDonald, Daniel Sharp, and Yuan Wu. Together, they launched my first adventures into philosophy, and gave me the sense that it can be a collaborative enterprise, and one that can be as fun as it is deep. I am especially grateful to Yuan, one of the most careful and patient philosophers I know. He has read and improved almost every part of this dissertation, especially in the early stages. Thanks are also due to the Berkeley graduate students who became both friends and interlocutors: Sophie Dandelet, Tyler Haddow, Jim Hutchinson, Jackson Kernion, Antonia Peacocke, and Kirsten Pickering. Thanks also to those who became friends while graduate students at other institutions, especially Dylan Bianchi and Rebecca Millsop.

xv Outside of academic philosophy, I want to thank Richard Cushman, whose high school English course gave words a new kind of life. He encouraged my eye to linger on, and draw connections between, a text’s most essential phrases. Without his guidance, I doubt I would have found my way to Hume and Kant. Just as important were the contemplative spaces offered by Yoshi Akiba of K¯ojinanZendo and Donna Christy of Far Leaves Tea. Yoshi-san taught me Japanese tea ceremony, and Donna gave me a home in a community of tea drinkers. Since then, tea has always been an anchor for me, allowing clear thought and conversation to flourish. Tea created many memorable discussions with my friends in high school and college, in which some early inclinations toward philosophy first sprouted up. I want to thank all those friends—you know who you are—but in particular Veronica Benjamin and Benjamin Brast-McKie, who themselves have chosen a philosopher’s life. For a lifetime of friendship, I am grateful to Josh Brett, Jordan D’Amato, Myla Green, Jonathan Goldblatt, Albert Kong, Brennan McCoy, and Antonio Jacques De Wolk. Finally, I am grateful to my family for always being there for me, and for being my most consistent conversational partners outside of the academy. Thanks to my late grandparents Gerson and Judith for showing that the logical and scientific can combine with the creative; to my aunts Mickey and Shaya for the most unconditional of love; to my parents Nat and Marilyn for always believing in me, and for raising me in a Jewish tradition of letters and interpretation; to my mother-in-law Julia for seeing me for who I am, and bringing me joy and laughter; and to my triplet brothers Ben and Sam for giving me my sense of an unshakable connection to a world shared with others—a connection which I suspect underlies my interest in talking to and curing skeptics. Most of all, I want to thank my husband and teacher, Eugene Chislenko, who has been my constant companion and friend. He has read every word I have written, and stood by my side, both literally and figuratively, throughout the whole production of this dissertation. I could never clearly separate where, in what follows, my ideas end and his begin, and vice versa. But that goes with the ‘with-life’ which we have chosen. Thank you, Zhen, for carrying me, and making the world more expansive, bountiful, beautiful, and real.

xvi 1.0 CURE AND PREVENTION

1.1 WHO SKEPTICS ARE

We can call someone ‘skeptical’ if she expresses doubt or suspicion in response to something she has just been told. Upon hearing, for example, that the California High-Speed Rail will be finished by 2027, she withdraws a little, tilts her head to the side, scrunches one side of her nose, and says, “I doubt that!” or “Can they really be so sure, after their past track record?” If you want her to trust that the plan will succeed, you can protest, “Don’t be a skeptic!” You could, if you wanted, call her a skeptic about the California High-Speed Rail Authority’s timeline. That would be a mouthful, but it would not be incorrect. People often become skeptical in response to the confident assertions of others. But the label ‘skeptic’ sticks for people who do not just react with fleeting doubt, but go on to tell others, perhaps repeatedly, that we cannot know certain things. Some such people wear the label as a badge of honor. There are skeptics of this sort about all kinds of things—about human-caused climate change, the deadliness of the corona virus, or the 1969 moon landing, to name a few. These skeptics claim that we cannot know something or other. They might go on to say that supporting evidence is, in one way or another, lacking. Such skeptics should be distinguished from deniers—people who deny, for example, that human activity is causing climate change. A climate change skeptic, in contrast to a climate change denier, denies that we know that human activity causes climate change, without necessarily denying that it does. It may be that skeptics easily slide into becoming deniers. But we can put this interesting bit of psychology to one side. My aim in what follows will be to talk to skeptics, whether or not they are also deniers. I will focus in particular on the kind of skeptics who have most interested philosophers. Such skeptics, whom we may call philosophical skeptics, have the following three characteristics. First, they claim that we cannot know, rather than that we do not. Their point is typically not that we should go on investigating, but that no investigation will ever lead to knowledge of the kind they deny we can have. Second, their skepticism is broad, targeting

1 an entire branch of human knowledge. They do not just claim that we cannot know about some particular event, object, or person, some scientiﬁc claim, or even some widespread phenomenon, such as climate change. Instead, they may claim that we can have no knowledge of how things are in the world around us, or of what is going on in the minds of other people, or of what happened in the past, or of causal laws, or of mathematics. Third, such skeptics put forward arguments in favor of their denial of our knowledge. Their arguments aim to base unobvious, skeptical conclusions on intuitively plausible premises which imply the conclusions. Presumably, such premises are meant to strike even those who are not yet skeptical as plausible, and so compel a common-sense thinker to become skeptical. In putting forward arguments, philosophical skeptics represent themselves as responding to more than a gut feeling. On the contrary, they tend to represent their conclusions as compulsory, or at least well grounded. Philosophical skeptics may then portray themselves as opposing dogmatists, who uncritically take for granted that they know all sorts of things. Going forward, I will focus on a particular kind of philosophical skeptic. This kind of skeptic claims that we cannot know how things are in the world around us, or, in other words, that we cannot know things about the familiar objects we ordinarily think we take in through perception, like rocks, trees, books, cities, dogs, people, and so on. We might call such a person a skeptic about perceptual knowledge, insofar as she denies that we can have the kind of knowledge we ordinarily take ourselves to acquire through perception. My reasons for focusing on this kind of skeptic are both historical and philosophical. Historically, this kind of skeptic has captured a great deal of attention from philosophers throughout the ages. Especially since the early modern era, the skeptic about perceptual knowledge has returned again and again in slightly variant guises. In what follows, I will discuss two of these guises: the external world skeptic (in Chapters1 and2), and the skeptical empiricist (in Chapters3 and4). For my purposes, both of these skeptics are skeptics about perceptual knowledge, though each comes into his or her skepticism in diﬀerent ways, and the latter’s skepticism may be viewed as attaching to more than just perceptual knowledge. I will say a bit more about the relationship between these guises of skepticism in §2.7 and Chapter5. Skepticism about perceptual knowledge is philosophically interesting in part because it

2 can be motivated in various ways without obviously relying on substantive philosophical assumptions. Perhaps even more interesting is the enormous breadth of the knowledge this kind of skepticism targets. Plausibly, we know much of what we know by perceiving it to be so, or on the basis of what we perceive. Through perception, we come to know all sorts of things about what shapes things are, where they are located, and what they look like. We know many of the ﬁndings of empirical sciences such as chemistry and epidemiology, and technical arts such as farming and construction, at least in part by making observations through perception. And, plausibly, we know things about the past by having previously perceived those things. We may perhaps even know things about what others are feeling by perceiving how they feel through or in their facial expressions and their gestures. If skepticism about perceptual knowledge is correct, we can have none of all of this supposed knowledge. Arguments for this kind of skepticism thus pose a threat to a great deal of human knowledge.

1.2 KINDS OF RESPONSES TO SKEPTICISM

Arguments for skepticism about perceptual knowledge pose a threat to much of what we take ourselves to know. If the skeptical arguments cannot be dispelled, or our claims to knowledge somehow defended, it seems we must admit that we cannot even know, by looking or feeling, such basic facts as that we have hands. We can attempt to respond to a skeptic in various ways. Jim Pryor, for example, distinguishes two kinds of responses to skeptics—or, as he puts it, two anti-skeptical projects—as diﬀering in degree of ambition. He says:

The ambitious anti-skeptical project is to refute the skeptic on his own terms, that is, to establish that we can justifiably believe and know such things as that there is a hand, using only premises that the skeptic allows us to use. The prospects for this ambitious anti-skeptical project seem somewhat dim. The modest anti-skeptical project is to establish to our satisfaction that we can justifiably believe and know such things as that there is a hand, without contradicting obvious facts about perception. This is not easy to do, for the skeptic can present us with arguments from premises we find intuitively acceptable to the conclusion that we cannot justifiably believe or know such things. So we have a problem: premises we find plausible seem to

3 support a conclusion we ﬁnd unacceptable. The modest anti-skeptical project attempts to diagnose and defuse those skeptical arguments; to show how to retain as many of our pre- theoretical beliefs about perception as possible, without accepting the premises the skeptic needs for his argument. Since this modest anti-skeptical project just aims to set our own minds at ease, it’s not a condition for succeeding at it that we restrict ourselves to only making assumptions that the skeptic would accept.1

Let us review Pryor’s distinction with regard to responses to skepticism which seek to establish that we can have knowledge, rather than justiﬁed belief.2 An ambitious response to skepticism tries to show that we can know what the skeptic claims we cannot know “using only premises that the skeptic allows us to use.” Such premises must not beg the question against the skeptic. For example, if the skeptic has put forward a plausible reason to think we can gain no knowledge through perception, it will not do to respond that we have some bit of perceptual knowledge. That would be to take for granted just what the skeptic has called into question. Showing that we can have the knowledge the skeptic claims we cannot have using premises which do not beg the question against him, in contrast, would “refute” the skeptic “on his own terms.” Since the skeptic “allows us to use” the premises, he would presumably accept them, and so be compelled to draw the conclusion that we can in fact know what he previously claimed we cannot. In drawing this conclusion, he would cease to be a skeptic. A person responding to the skeptic in the ambitious way, then, talks to the skeptic and tries to convince him that he should not remain a skeptic. A modest response to skepticism also tries to show that we can know what the skeptic claims we cannot. But this response is not constrained to using only premises that do not beg the question against the skeptic. This is because its aim is not to convince a skeptic, but only to “set our minds at ease.” Pryor’s ‘our’ here refers to a community of common-sense thinkers, who take themselves to know what the skeptic claims we cannot. The way the modest response attempts to show “to our satisfaction” that we have this knowledge is to “defuse” any skeptical arguments which seem to call it into question. This is “not easy” because skeptical arguments to the eﬀect that we can have no perceptual knowledge tend to employ “intuitively acceptable” premises about perception. Insofar as the premises are “intuitive,” they can appear to express “obvious facts” or “pre-theoretical beliefs” about

1Pryor(2000), 517–18. 2For an explanation of why I focus on knowledge, rather than justiﬁed belief, see the Preface.

4 perception. So defusing the skeptical argument may take the form of showing that its premises are not really the obvious facts they can seem to be. This may involve slight amendments to any pre-theoretical beliefs about perception on which the argument relies, but these amendments are minimized. Insofar as the response is modest, it need not attempt to convince the skeptic that her premises are false or not properly representative of the nature of perception. The response succeeds as soon as it allows common-sense thinkers to “satisfy” themselves with the thought that the skeptic’s argument does not go through. The modest response is thus a response to the skeptic only insofar as it occurs after the skeptic has articulated an argument which makes a defense of common-sense knowledge necessary. But it is not a response to the skeptic. It is instead addressed only to fellow common-sense thinkers. A person responding to the skeptic in Pryor’s modest way, then, listens to the skeptic. But he does not speak back. Instead, he turns around and starts a conversation with only those who already think like him. The labels ‘modest’ and ‘ambitious’ are less than ideal for capturing the distinction Pryor has in mind. The ambitious response is more “ambitious” than the modest one in the sense that it seeks not only to satisfy like-minded, common-sense thinkers, but to convince skeptics as well. But the ‘modest’ and ‘ambitious’ labels clash with the thought that both kinds of responses can vary in ambition in other ways—for example, concerning the number of skeptical arguments either response is responding to. That suggests there can be modest responses which are more ambitious than certain ambitious responses, and vice versa. Pryor’s labels can obscure this possibility. If the difference in ‘ambition’ can be spelled out in terms of the breadth or number of people whom the response addresses, a different set of labels may be better suited to capture the distinction. The labels ‘modest’ and ‘ambitious’ muddy the waters in another way as well. For they are equally apt to capture another kind of distinction which Pryor can seem to be after. This is a distinction between different strategies or methods through which one might try to resist skepticism. When Pryor introduces the ‘ambitious’ response, he introduces it as seeking to “establish that we can. . . know such things as that there is a hand.” And though he says the same thing of the modest response, he goes on to clarify that the modest response seeks to do this by “defusing. . . skeptical arguments.” In other words, the ambitious response can

5 seem to require a direct argument to the effect that we can know what the skeptic claims we cannot, while the modest response can seem to opt for a different method or strategy— namely, showing that the skeptic’s argument is flawed and does not go through. Attempting to prove that we can know something can seem more “ambitious” a task than showing that one argument against our knowledge fails.3 It can then seem that Pryor’s distinction is a distinction between methods or strategies for resisting skepticism, rather than a distinction between the intended audiences of the responses. His labels allow this ambiguity to thrive. This is unfortunate not only for the unclarity it creates, but also for the false impression it gives that the distinctions in strategy and audience overlap perfectly when in fact they do not. One may try to convince a skeptic that his argument fails on his own terms. And one may try to prove to her own satisfaction that she can know what the skeptic claims she cannot, even if this begs the question against the skeptic. In other words, one can be ‘ambitious’ in strategy but not audience, or vice versa. I propose, then, to introduce new labels to better capture the two different distinctions. The first distinction, which concerns the intended audience of the response to skepticism, characterizes mutually exclusive kinds of responses. A response cannot be meant to appeal to everyone, including skeptics, and at the same time be meant to appeal to only non- skeptical, common-sense thinkers. It is probably this distinction that Pryor is after, given the contrast between such phrases as “on [the skeptic’s] own terms” and “to our satisfaction,” “using only premises the skeptic allows us to use” and “not. . . restrict[ing] ourselves to only making assumptions that the skeptic would accept.” Following Timothy Williamson’s medical metaphors, I will describe this distinction as one between the “cure” and “(mere) prevention” of a skeptical condition.4 A curative response, then, seeks to convince a skeptic that we can know what she thinks we cannot; it seeks to cure her skepticism. A (merely) preventive response, on the other hand, attempts to show only common-sense thinkers how

3This reading of Pryor’s distinction can be further encouraged by the use of the ambitious/modest distinction in the literature on transcendental arguments. An ambitious transcendental argument concludes with a claim about non-psychological reality. Presumably, a sound ambitious transcendental argument could then secure our claim to some bit of knowledge which a skeptic doubts. See Brueckner(1996), 265–66 and Stroud(1999), 160–61. 4Williamson(2000), 26–27. I quote a passage in which Williamson uses this distinction on p.9 below.

6 to avoid being taken in by skeptical arguments; it thus seeks only to prevent new onsets of skepticism, not to cure those afflicted with it. We can continue to use medical metaphors to capture the other distinction latent in Pryor’s discussion. We can call a response to skepticism supplemental if it aims to argue that we can or do know what the skeptic claims we cannot. The response is supplemental both in the sense that it responds to the skeptic with a supplemental argument for our knowledge and in the sense that this argument serves as a supplement which shores up our health against a skeptical condition. We can call a response to skepticism antidotal if it aims to neutralize the skeptic’s argument against our knowledge, by showing that the argument fails or is insufficiently supported. This response is antidotal in the sense that it counteracts the poison of the skeptical argument. These two responses are not mutually exclusive and are sometimes combined. One can, for example, try to show that a skeptical argument fails because some of its true premises imply or presuppose the very knowledge it calls into question.5 Once we have recast Pryor’s distinctions in these ways, a further clarification is in order. It is sometimes possible to change someone’s mind without arguing using only premises and inferential steps that are consistent with and acceptable from her initial viewpoint. That might at first sound incredible. But it in fact happens all the time in ordinary conversation. In a quarrel with a partner or spouse, you may say, “Here is how I see what happened.” Even if what you describe is at odds with how your partner experiences the cause of the quarrel, and so is, in a sense, unacceptable to them, they might nonetheless be able to step back and try to see things from your perspective. Doing this may soften their outlook and even change their mind, all while your telling of what happened, in some sense, begs the question against their initial view. Something similar can occur in philosophy as well: We can change someone’s mind by sketching an alternative way of looking at things. This alternative can, in some circumstances, appeal to the person one is talking to even while it is inconsistent with, or even fundamentally opposed to, her initial views. It is easy to overlook this dialectical move if we too strongly or narrowly conceive of a cure for skepticism as “using only premises and inferential steps the skeptic will accept.” It will be useful for keeping our

5See Stern(2000), 5.

7 distinctions sharp, then, to shift the emphasis away from the resources a response uses and to focus instead on whether the response aims to change the skeptic’s mind. Of course, these aspects of cure cannot come apart entirely; it only makes sense to attribute the aim of changing a skeptic’s mind to a response to skepticism when that response shows some indication of trying to understand what considerations can actually appeal to a skeptic. All I am suggesting here is that we should not think that only claims and inferences consistent with the skeptic’s position could ever appeal to him. This clariﬁcation will prove to be important for the cure for skepticism I discuss in Chapter4. We now have three sharper distinctions between kinds of responses to skepticism: cure and prevention, supplement and antidote, and responses that do and do not appeal only to premises that are consistent with and acceptable from the skeptic’s initial viewpoint. Having these distinctions in hand will help us better understand the prospects for each kind of response.

1.3 THE PREFERENCE FOR PREVENTION

With a clariﬁed distinction between cure and prevention, we can now articulate a current trend in the literature. There is an increasingly widespread attitude among epistemologists that what I am calling the curative response is doomed. According to Pryor, for example:

The prospects for [the] ambitious anti-skeptical project seem somewhat dim. . . Most falli- bilists concede that we can’t demonstrate to the skeptic, using only premises he’ll accept, that we have any perceptual knowledge. . . . [T]he ambitious anti-skeptical project cannot succeed.6

Pryor opts, instead, for a modest response. He urges common-sense thinkers to adopt what he calls a “dogmatist theory” of perception. This theory, he claims, undermines a premise in his opponent’s skeptical argument. But he does not oﬀer the skeptic any reasons to adopt

6Pryor(2000), 517–20. Compare Pryor(2004), 370: “Nowadays, it’s commonly agreed that an adequate philosophical response to the skeptic need not be capable of rationally persuading the skeptic that the external world exists, or that we have justiﬁcation to believe it exists. Nor need it be capable of persuading someone who’s seized by skeptical doubts. What it does have to do is diagnose and explain the ﬂaws in the skeptic’s reasoning. It has to explain away the intuitions that the skeptic draws support from.”

8 his dogmatist theory. For Pryor, trying to change the skeptic’s mind is futile.7 How does Pryor think the modest strategy can succeed? Let us grant that Pryor is right that if the dogmatist theory is true, the skeptical argument fails. Pryor says he motivates his dogmatist theory using a “standard philosophical methodology” of “sensible philosophical conservativism”—namely, to “start with what it seems intuitively natural to say about perception, and. . . retain that natural view until we find objections that require us to abandon it.”8 He claims that common-sense thinkers will find his dogmatist theory of perception intuitively plausible. The skeptic, in contrast, does not. Her sense of what is intuitively plausible has been transformed as a result of her skeptical thinking. She may then raise an objection to the dogmatist theory in the form of a skeptical argument. But Pryor does not think the skeptic’s objection actually requires common-sense thinkers to give up what they find intuitively plausible to think about perception. That is because holding the dogmatist theory protects common-sense thinkers from taking on a premise in the skeptic’s argument. Having grounds for rejecting that premise, then, allows the common-sense thinker to dismiss the skeptic’s objection, and so hold on to the theory. For the skeptic, however, those grounds have been shaken and cannot be restored. Hence, Pryor thinks an appeal to the dogmatist theory can only prevent skepticism, not cure it. Similarly, Williamson offers a way of thinking about knowledge which, if adopted, “blocks” “one route into scepticism.” But he does not seek to offer a way back. He says:

If someone has already taken the route into scepticism oﬀered by [a] fallacious [skeptical] argument, before it was blocked, and has become genuinely undecided, at least in principle, as to whether she is in a sceptical scenario [i.e., a scenario in which she can form no knowledge of the world around her],9 then the blocking of the route now comes too late to rescue her. Nothing said here should convince someone who has given up ordinary beliefs that they constitute knowledge, for nothing said here should convince her that they are true. The trick is to never give them up. This is the usual case with philosophical treatments of skepticism: they are better at prevention than at cure. If a refutation of skepticism is supposed to reason one out of the hole, then skepticism is irrefutable.10

Williamson claims that his conception of our knowledge allows common-sense thinkers to

7Pryor(2000), 532, 534–35. I discuss the targeted skeptical argument in §2.2 and criticize Pryor’s modest response in note 27 of that chapter. I develop a criticism of modest responses in general in §§1.4–1.5. 8Pryor(2000), 538. 9Skeptical scenarios are introduced and discussed in §2.2 below. 10Williamson(2000), 26–27, my emphasis.

9 resist a premise in a skeptical argument, which is thus shown to be “fallacious.” But he does not think his conception can help someone who has already become convinced by the skeptical argument. Such a skeptic, he thinks, must be given a story about how the kinds of beliefs we ordinarily form on the basis of perception—for example, that there is a bird on the branch outside the window—are true. His account of our knowledge does not offer such a story. And he thinks that no story can suffice once a skeptic has become convinced that she may well be in a situation in which all of what she seems to perceive is illusion. Prevention rather than cure, he concludes, is the only viable response to skepticism. Many others now agree.11 In Chapter2, I will argue that prevention is not the only response to skepticism left open to common-sense thinkers. I will do so by sketching a cure for skepticism about perceptual knowledge, there in the guise of skepticism about the external world. Here, I want to consider the viability of the preventive response. Against the current trends, I think it is this response whose prospects are “dim.” The rest of this chapter is devoted to defending this prognosis. My argument will be general in form. I want to draw suspicion toward the idea that a response which cannot reach someone convinced of skepticism can succeed at, as Pryor puts it, “set[ting] our own minds at ease.” In putting forth this argument, then, I will not engage with particulars of the preventive responses we find in Pryor, Williamson, or any other philosophers. For all I say here, it may very well be that Pryor’s dogmatist theory of perception, or Williamson’s conception of our knowledge, can figure in a successful response to skepticism. But if either in fact does, I think they must be able to be made acceptable to the skeptic, and so figure in a cure rather than a mere prevention. My argument will begin with an observation that I think even proponents of the preventive strategy agree with. Proponents of that strategy are generally comfortable with

11See Nozick(1981), 15–16; Burge(1993), 548–49; Byrne(2004), 300–03; Peacocke(2004), 65, 70; Wright (2004), 174–75; Enoch and Schechter(2008), 554; Ginsborg(2011), 240–41; Boghossian(2012), 47; Byrne (2014), 278, 285; DeRose(2017), 68–69, 84–86. For recent exceptions, see Greco(2012) and Rinard(2018). In treating the skeptic as unreachable, proponents of the merely preventive response do not seem to be treating skeptics as responding irrationally. This suggests that such proponents may have a commitment to epistemic permissivism, the view that rationality allows multiple permissible responses to a consideration or body of evidence. This thesis is controversial, and rarely articulated by such proponents. For the thesis, see White(2005). For an overview of arguments for and against, see Kopec and Titelbaum(2016). For a more recent argument against, see Schultheis(2018).

10 begging the question against the skeptic. But I think they would admit that not every kind of question-begging is on equal footing. Some valid, question-begging arguments to the effect that we can know what the skeptic claims we cannot know seem to miss the skeptic’s point entirely. And those arguments can hardly be thought to ease the mind of someone who has really heard what the skeptic has to say. In §1.4, I will discuss one such argument: G.E. Moore’s notorious proof of the external world. This discussion will lead to the moral that responses to skepticism must seek to understand why the skeptic’s argument can seem to be plausible, if they are to succeed. This alone does not show that the preventive response fails. But, as I argue in §1.5, it puts the preventive response into a dilemma: For a preventive response to do more than just ignore the skeptic, it must address the ways in which the skeptic’s argument appears compelling. In doing this, a person responding to skepticism may either find that the skeptical argument is compelling, in which case she would find skepticism compulsory and would herself stand in need of cure; or she may come to see how the argument is not compelling, in which case the results of her inquiry would constitute a cure. If that is correct, a preventive response which does not just ignore the skeptic either fails on its own terms or transforms into a cure. Either way, there is no room for a successful merely preventive response.

1.4 IGNORING THE SKEPTIC

A skeptic about perceptual knowledge may express his view by saying that we know nothing about things external to our thought. Presumably aware of such skeptics, Moore argued that there are external things. His proof went like this: “Here is a hand, here is another, therefore there are external things.”12 (We may imagine Moore’s accompanying hand gesticulations.)

12Moore ﬁrst expresses his proof quite generally: “I think. . . that in the case of all kinds of ‘things’, which are such that if there is a pair of things. . . then it will follow at once that there are some things to be met in space, it is true also that if I can prove that there are a pair of things. . . then I shall have proved ipso facto that there are at least two ‘things outside of us’; if I can prove that there exist now both a shoe and a sock, I shall have proved that there are now ‘things outside of us’, etc.; and similarly I shall have proved it, if I can prove that there exist now two sheets of paper, or two human hands, or two shoes, or two socks, etc.” Moore then goes on to give the most memorable instance of such a proof: “I can prove now, for instance, that two human hands exist. How? By holding up my two hands, and saying, as I make a certain gesture

11 It is easy to feel that this proof is unsatisfying. But it is worth pausing to note, as Moore does, that the proof has a lot going for it. It does not obviously fall into the vice of circularity, insofar as the conclusion, that there are external things, could be true even if the premises were false—that is, even if there were no human hands here, or only one. Plausibly, the premises are known; what better way is there to know that a hand exists than to see it before one’s eyes? The premises seem to imply the conclusion, since it seems a hand is an external thing, if anything is. Since, plausibly, the proof is non-circular, its premises are known, and the conclusion follows validly from its premises, anyone who grasps the proof can learn its conclusion. Moore’s proof would provide that person with knowledge that there are external things.13 The possession of this knowledge would directly contradict the skeptic’s denial that we can have it, and indeed do so using the same terms that she is apt to employ in expressing her skepticism. Moore’s proof is sound and is free from formal and logical vices.14 It provides us with knowledge of what the skeptic doubts, and so provides us with a counterexample to the skeptic’s view that we can know nothing about external things. What then goes wrong with Moore’s proof? Many have claimed that the proof fails because it “begs the question” against the skeptic.15 It seems to do this by assuming such things as ‘here is a hand.’ We are to imagine Moore holding his hands up for the audience of his proof to see, while he gives it. And so we are to imagine the premises to be known to be true by Moore and us by our having seen his hands before our eyes. But the skeptic concludes that we can never know

with the right hand, ‘Here is one hand’, and adding, as I make a certain gesture with the left, ‘and here is another’. And if, by doing this, I have proved ipso facto the existence of external things, you will all see that I can also do it now in numbers of other ways” (Moore, 1939/1993, 165–66). 13On the virtues of Moore’s proof, see Moore(1939/1993), 166 and Stroud(1984), 84–86. 14Some have argued that the proof, though valid, nonetheless suﬀers from a formal problem, roughly because entitlement to the conclusion is presupposed by entitlement to each premise. See Wright(2002), 331–33, 336–37. An implication of this charge is that we can only know each premise—for example, that here is a hand—if we already know the conclusion—that there are external things. If that is right, the premises cannot be used to support the conclusion. But the charge is false. Even if it is true that entitlement to the claim ‘here is a hand’ depends on prior entitlement to some general expression of our perceptual entitlement (which would require some argument to show), it is not clear why it would depend on perceptual entitlement to multiple external things. A hand is only one external thing. 15Pryor(2000), 518; Wright(2002), 330, 332, 337; Carter(2012), 116. Some have levied this criticism not against the proof itself, but only against straightforward readings of it. See Ambrose(1952), 399; Malcolm (1952), 348–49; Malcolm(1963a), 177. For a reply that Moore’s proof should be read straightforwardly, see Stroud(1984), 88–105. Many charge anti-skeptical strategies inspired by Moore’s proof with question- begging. See Wright(2002), 344ﬀ; Pritchard(2008), 284–85.

12 such things as ‘here is a hand’ by seeming to perceive a hand before our eyes. And Moore does not give the skeptic a reason to think we do or can have such knowledge. Thus, Pryor concludes: “This is why Moore’s ‘proof’ strikes us as so unsatisfactory: he hasn’t offered any non-question-begging reasons to believe his premises.”16 It is worth going slowly with this charge of question-begging. It is not as if Moore’s proof begs the question against everyone just because he has not given reasons to believe his premises. Someone who has never encountered a bit of skeptical reasoning will have no problem accepting Moore’s premises, without needing to hear more about why those premises are true. Such a person may perhaps find Moore’s proof eccentric or pointless, lacking the context of the threat posed by skeptical arguments. But she would not find the proof question-begging. The skeptic does, since she has argued against our coming to know things through perception. We might then be tempted to say that Moore’s proof begs the question because it makes assumptions, knowledge of which the skeptic concludes we cannot have. But the problem cannot just be that the proof takes for granted knowledge which is inconsistent with, and so challenged by, what the skeptic concludes. In general, it is often legitimate to reject uncertain claims on the grounds that they conflict with things one takes oneself to know. Say that Stephen tells me that Vivian is in the office, because he saw a steaming cup of tea on her desk. If she just told me that she would be stepping out to run a sudden errand, I can report that she is out running errands, and so dismiss Stephen’s claim as inconsistent with what I take myself to know. I would not be doing anything foolish or illegitimate by not investigating further into the reasons why what I say might be true.17 Indeed, in dismissing what Stephen concludes from seeing the cup of tea, I would not beg the question against him. This is so despite the fact that, if what he concludes is true, I do not know what I take myself to know. If I do not beg the question against Stephen, there must be some further reason why Moore, in claiming that here is a hand, begs the question against the skeptic. It is not just

16Pryor(2000), 518. 17As Stroud(1984), 105 notes, “dismissing without further investigation something that conﬂicts with what is already known is the very heart of rationality.” For a helpful discussion of this point with regard to Moore’s proof, see 103–12.

13 that he relies on knowledge which the skeptic does not think he has. After all, I rely on knowledge which Stephen is committed to my not having when he claims that Vivian is in the office. But say Will tells me that he ran into Vivian a few minutes ago sorting some papers from her mailbox. He goes on to tell me that she decided to run the errand tomorrow. If I respond to Will that he cannot be right because Vivian has just stepped out to run an errand, I would be doing something foolish. I would beg the question against Will. The reason why I beg the question against Will but not Stephen is that Will’s reason for thinking Vivian is in the office undermines my reason for thinking she is not. If I do not recognize this, I am either blatantly ignoring what Will has said or I am a fool. We might want to say the same of Moore. The skeptic puts forward an argument, using intuitively plausible premises, that skepticism is true. If the skeptic’s premises really are plausible, and her argument appears sound, she has given Moore reasons to think that he cannot know things by perceiving them to be so. In responding to the skeptic with premises, knowledge of which the skeptic has given plausible reasons to think we cannot have, he seems either to be just ignoring the skeptic’s argument or to be a fool. Moore was not a fool. It is true that his missteps have long interested epistemologists, often in the guise of an instructive, cautionary tale.18 I continue in that tradition. At the same time, Moore has also been celebrated as a champion of common sense. He has appeared to some to be heroic in his unwillingness to take on premises in the skeptic’s argument. But to others, this has made his quarrel with the skeptic hard to understand. In positioning himself against the skeptic, Moore seems to portray skeptical arguments as constituting threats to our knowledge.19 And that suggests that he finds them to have at least a degree of intuitive plausibility.20 But, as soon as he responds to the skeptic with, “Here is a hand,”

18See Stroud(1984), 100, 126: “Moore does something of deep philosophical interest” despite the fact that “what Moore says is no refutation of philosophical scepticism.” Williams(1991), 43: “although we are all certain that the sceptic cannot be dismissed in [Moore’s] way, it is not entirely easy to say why not.” Pryor(2004), 349: “ Something about this argument sounds funny. As we’ll see, though, it takes some care to identify exactly what Moore has done wrong.” 19This is suggested not only by his “Proof of the External World,” but also by his continuous interest over a roughly thirty year span in overturning principles he saw as leading to skepticism. For a discussion of the relevant lectures and texts, see Stroud(1984), 103–08. 20Sometimes, Moore even seems to accept skeptical principles which are in clear tension with his comfortable reliance on particular instances of perceptual knowledge. For example, in Moore(1941/1993), 191, he says of the argument “since I do know that I’m standing up, it follows that I do know that I’m not dreaming” and the argument “since you don’t know that you’re not dreaming, it follows that you don’t know that you’re

14 it is as if he has never considered her argument to count against our coming to know things through perception. He seems to have forgotten the argument entirely, or to have decided to ignore it. And in the same breath, he seems to have forgotten who he is talking to, or to have decided to ignore that she is a skeptic. By reverting to the common-sense outlook in the middle of a quarrel with a skeptic he pretends to take seriously, Moore does something which is hardly common. As Barry Stroud puts it, “If Moore in his response represents the ordinary man, he is a most extraordinary ordinary man in not being ‘lured’ into the traditional philosopher’s understanding of his question.”21 At this point, we may wonder if Moore’s proof has seemed to fail only because we have expected that it must cure the skeptic. After all, in spelling out what seems dissatisfying about it, I have focused on the ways in which it begs the question against skeptics. But, as I have explained, many philosophers now think that responses to skeptics can succeed, even while begging the question against the skeptic. I have called this kind of response (mere) prevention. Does Moore’s proof fare any better, if reconsidered as a kind of prevention? I think the answer is no. Many people who feel a dissatisfaction reading Moore do not consider themselves to be skeptics. But they do find skeptical arguments to appear plausible, at least before a close inspection of all the considerations they employ. Moore seems to beg the question against these readers just as much as he does against the skeptic. Even proponents of the preventive response admit this. After all, Pryor says that “Moore’s ‘proof’ strikes us as. . . unsatisfactory.”22 Presumably, it would thus fail to “set our minds at ease” as common-sense thinkers. For Pryor, the preventive response must resolve a “problem” posed by the skeptical argument—namely, that “premises we find plausible seem to support standing up” that “the one argument is just as good as the other.” If the latter argument really is just as good as the former, skepticism is not far away. See my §2.2. Again, Moore seems to acknowledge the force of a related skeptical idea in Moore(1939/1993), 169, when he says: “How am I to prove now that ‘Here’s one hand, and here’s another’? I do not believe I can do it. In order to do it, I should need to prove for one thing, as Descartes pointed out, that I am not now dreaming.” I do not mean to suggest that Moore thinks such a proof is required for his knowing ‘here is a hand.’ He does not. Instead, I cite this passage as evidence that Moore is not oblivious to the ideas which get the skeptic’s argument going. He sometimes accepts them wholesale, and at other times in at least a conditional way; in the latter case, he accepts that if we must prove that ‘here is a hand’ to know it, we would need to show first that we are not dreaming. If Moore did not find the skeptical arguments at least somewhat plausible, it is unclear why he would accept even this conditional claim. 21Stroud(1984), 199. See also 124: “Moore takes general questions and assertions expressed in the same words as those of philosophers and answers or refutes them by appeal to particular examples.” 22Pryor(2000), 518, my emphasis.

15 a conclusion we ﬁnd unacceptable.”23 Moore’s response does not do that. In fact, it does not go so far as to consider why the skeptic’s premises appear plausible, let alone explain how this plausibility can be shown to be illusory. Instead, Moore just seems to ignore the whole predicament.24 The dissatisfaction that even common-sense thinkers feel about Moore’s proof suggests a general moral for any response to skepticism, whether curative or merely preventive: For a response to succeed, it must engage with the considerations which seem to make a skeptical conclusion appear compulsory. And it must give us some way to drain those considerations of their apparent intuitiveness. Otherwise, we will be stuck with the “problem” which Pryor’s merely preventive response is meant to address. Such a response would beg the question not only against the skeptic, but also against any common-sense thinker who views skeptical arguments as in any way worthy of response. For a response to skeptics to succeed, it then seems, it should pay attention to skeptical arguments and seek to understand their appeal.

1.5 A DILEMMA FOR MERE PREVENTION

Pryor and others admit that their responses to skepticism will beg the question against the skeptic. Are their responses any better than Moore’s? The quick answer seems to be yes, insofar as they consider and try to disarm skeptical arguments; Pryor and others claim to give us reasons to reject certain premises in the skeptical arguments they aim to disarm. But in the end I think that, like Moore, they fail to remove the bite of the skeptic’s arguments. In portraying their responses as merely preventive, they refuse to investigate what leads the

23Pryor(2000), 517, my emphasis. 24Another issue with Moore’s proof, conceived as an anti-skeptical argument, is that its appeal to multiple hands is entirely superfluous. By this, I mean that knowledge of either of its premises already provides a counterexample to the skeptic’s denial of our knowledge. A skeptic may very well express her skepticism about perceptual knowledge in terms of a denial that we can know of external things. But knowledge of any one external thing, such as a hand, is already knowledge which the skeptic denies. Appealing to knowledge of multiple external things, as Moore does when he expresses his conclusion (see n. 12 above), adds nothing of substance. So, even though the proof is non-circular, it already accomplishes all that it can be thought to with regard to skepticism in the statement of its first premise. The same can be said of the second premise. As Wittgenstein(1969/1972), §1 puts it: “If you do know that here is one hand, we’ll grant you all the rest.” I suspect that this issue is what inclines us to charge the proof with a logical flaw of circularity, which does not in fact fit.

16 skeptic’s premises to appear plausible. I think the problems with their particular attempts to prevent skepticism are instances of a more general problem with any merely preventive response. Let us grant that I am right in the previous section: that for a response to skepticism to succeed, it cannot just ignore the considerations leading to skepticism, but must instead seek to understand why they have seemed plausible. If we are thorough in our attempts to understand this, we will find one of two things. We might find that the considerations in favor of skepticism are indeed plausible. Or we might find that those considerations only appear plausible, but can be shown not to be. Preventive responses then face a dilemma. In the first case, after spending some time with the arguments, the person responding to skepticism finds them to be plausible. She now finds herself with some reasons to doubt that she can know what the skeptic claims she cannot know. The skeptical argument may not have transformed her into a philosophical skeptic, who insists and argues that we cannot know. But she is now moved by the same considerations that move that skeptic. At very least she now feels a kind of perplexity about our knowledge. Removing this perplexity would require the same kind of cure that could change a philosophical skeptic’s mind, given that she is in the grips of exactly the same considerations which seem to support the skeptic’s position. She at least partially shares the skeptic’s point of view, and so stands in need of a response which can be convincing to someone who sees things as the skeptic does. At this point, prevention would come too late. In the other case, the person responding to skepticism has found that the skeptic’s considerations only appear plausible, but can be shown not to be. She then has a way to undermine the plausibility of those considerations. She may, for example, be able to show that certain premises in the skeptical argument appear plausible only under further assumptions which are not intuitive and would require significant argument to support. Any such findings about the skeptical argument’s structure would presumably be available to the skeptic. The person responding to the skeptic would then be well on the way to developing a cure, as those conditions or assumptions could be communicated to the skeptic. The development of this cure would remove the motivation to settle for mere prevention. Note that I have not argued that any prevention of skepticism must fail. Instead, I have

17 argued that any successful prevention must also be a successful cure. In fact, a cure which disarms skeptical arguments will always also be prevention, because it drains the plausibility of an argument which could otherwise lead non-skeptics to become skeptics. On the other hand, we now have some reason to doubt that there can be a successful merely preventive response. Attempts at prevention will lead either to illness, requiring a cure, or to the cure itself. In posing the dilemma, I have assumed that skeptical arguments are at least initially appealing and worthy of a close inspection. In doing this, I am in good company.25 But some skeptical arguments can be dealt with without an exploration of their appeal. For example, when they are based on a blatant error, such as a basic logical fallacy or a typo, we can dismiss them without consideration of the appeal of their premises. Presumably, the presence of such blatant errors could be communicated to the skeptic who makes the argument. This would be a cure which even proponents of the preventive response would acknowledge. But these proponents tend to view skeptical arguments as typically more plausible, and thus requiring something like “diagnosis” and “defusing.” Here, one might object as follows: Do the preventive responses of Pryor and others not already avoid my dilemma, by “defusing” skeptical arguments using considerations that only non-skeptics find plausible? Are these not successful, merely preventive responses? I think not. Pryor’s response to a skeptical argument is meant to provide the grounds for rejecting a premise in that argument. He suggests that common-sense thinkers adopt a dogmatist theory of perception which is inconsistent with one such premise. Presumably, adopting this theory could make the premise appear less plausible. In suggesting a way to drain the premise’s plausibility, Pryor does not think he is offering a cure, because he thinks that the skeptic will not find his dogmatist theory plausible. He argues for that theory on the grounds that it accords with pre-theoretical, common-sense beliefs about perception which he thinks the skeptic no longer shares. But the reason the skeptic no longer believes these things is that her skeptical argument calls them into question. If that is right, it is not clear how Pryor himself can appeal to such common-sense beliefs as support for his theory before he has already drained his targeted premise of its plausibility. If Pryor takes for granted that

25See Chapter2, n.6.

18 the premise no longer appears plausible in adopting his dogmatist theory, he seems to be doing just what Moore did. He seems to be ignoring or forgetting the skeptical argument. But if he recognizes the premise to be plausible, then he can no longer help himself to the grounds on which he adopts that theory. He should instead recognize his own need for cure. Finally, if he thinks the grounds remain available even while viewing the skeptic’s premises as plausible, then those grounds should in principle be available to the skeptic as well. In that case, his “anti-skeptical strategy” would be more than preventive.26 Though I will return to Pryor in Chapter2, we do not at this point need to see in which of these ways Pryor understands his own response. It is enough to see that the dilemma applies. Any response to skeptics that does not ignore their arguments must either acknowledge or remove their plausibility. We can see here that the dilemma is fairly general. Whenever an argument appears sound, there is a strong pressure toward accepting its conclusion. There may be resources for resisting the argument that are available only to those who do not accept its conclusion. But, to the extent that the argument grips us, we cannot appeal to those resources ourselves. I took as a moral from our dissatisfaction with Moore’s proof that skeptical arguments do grip us. If the same can be shown of an argument in any philosophical debate, or indeed outside of philosophy, it seems there is little room for the analogue of mere prevention. If we feel the pull of an apparently sound argument, it would be diﬃcult to remove that pull by appealing to resources whose availability depends on one’s not being taken in by it. The

26A comparison to Pryor’s discussion of Moore’s argument is helpful here. Pryor(2004), 369 says: “In the happy case where you neither have nor have reason to have the kinds of doubts that the skeptic wants to induce, then the justification your experiences give you for Moore’s premise [‘here is a hand’] will be undefeated and unobstructed. Having that justification for the premise will make Moore’s conclusion more credible for you; and that justificatory relationship is one that you can rationally endorse in your reasoning.” But, he continues: “Clearly Moore’s argument is not very dialectically effective against [an external world] skeptic. . . The skeptic has doubts that prevent Moore’s argument from rationally persuading him.” For Pryor, Moore’s argument will only constitute a defense of common-sense perceptual knowledge, to someone who has not been taken in by skeptical doubts, and has not been given a reason to take on those doubts. The same can be said of Pryor’s assumption of a dogmatist theory, insofar as that theory is adopted in virtue of its accordance with pre-theoretical views about perception which he thinks the skeptic does not share. But the ease with which we feel dissatisfied with Moore’s argument suggests that until the skeptic’s argument has been disarmed, it does induce the relevant doubts, or at least presents us with a reason to take on those doubts. If Pryor does more than just ignore the skeptic’s argument, and if Pryor is right that the grounds for the dogmatist theory are unavailable to someone in the grip of doubts, it is unclear how Pryor can adopt a dogmatist theory before he has already disarmed the argument. So it is unclear how he can adopt the theory as a means to disarming the argument.

19 seeming soundness of the argument makes those preventive resources seem unavailable. This difficulty can motivate the search for a way to disarm the argument which is palatable even to those who already accept its conclusion. Another objection is worth putting to rest. The dilemma can seem to imply that whenever we find we cannot cure a skeptic, we must ourselves be perplexed and in need of cure. This seeming implication seems implausible. Surely skeptics can be stubborn and resist cure, even when it is a proper cure for them. But I do not think the dilemma has the implausible implication. When the skeptic is stubborn, the person responding to the skeptic can still have developed a way of avoiding the perplexity that results from finding the skeptical argument plausible. This way of avoiding the perplexity can successfully change the minds of skeptics who consider it on its merits. The fact that some skeptics refuse to listen does not make a cure any less of a cure. Someone who refuses to take antibiotics to treat an infection does not thereby show that the antibiotics are not a cure. Nor do most known medical cures cure 100% of those to whom they are administered. We should distinguish having or offering a cure, curing someone, and curing everyone. It is then easier to see why we might fail to cure some skeptics, without being perplexed ourselves, doubting that we have a genuine cure, or opting only for prevention. Like some medical patients, some skeptics may be incurable. With the preceding discussion in mind, we can identify three powerful motivations for developing a cure for skepticism. First, even if we cannot cure every skeptic, developing a cure is a humanitarian concern. It allows us to help people who are unable to accept the possibility of knowledge in large and crucial areas of human concern or activity. Second, we may be able to learn from developing a cure. Insofar as skeptical arguments seem to show us something deeply perplexing about our claims to knowledge, engaging more deeply with them is educational, promising to teach us something about ourselves and our knowledge. Third, since we have found that mere prevention is unlikely, a cure is our best chance at addressing the threat posed by skeptics and their arguments. Cures thus have a preventive aim as well. In most of the rest of this dissertation, I will develop a cure for skepticism, driven by all three of these motivations. As we will see, I also think that we can learn something from listening to skeptics, apart from curing them.

20 1.6 TRANSITION: HOPE FOR A CURE

In considering the dilemma, I have sketched what one possible kind of cure can look like. This curative response sympathetically considers skeptical arguments and asks on what conditions, or under what assumptions, their premises appear plausible. If it can be shown that those conditions are optional, or that the premises rely for their plausibility on a commitment to unobvious and unsupported assumptions, showing this would constitute the heart of a cure. Such a finding would enable a skeptic to see that he need not draw the conclusion of his skeptical argument, and would remove his sense that our knowledge is impossible. A complete cure of the skeptic may perhaps still require a further explanation of how we can know what the skeptic previously argued we cannot know. But a skeptic who has given up his denial of our knowledge may very well take an interest in this explanation. And with the skeptical argument disarmed, ordinary considerations about perception can once again be brought in to support the explanation of our knowledge. We can call a method of response along these lines diagnostic, insofar as it searches for the reason why skepticism appears plausible, and seeks to remove this root cause. A possible result of a diagnostic response is that, in identifying a root cause, it may allow us to view various aspects of skepticism to be symptoms of the condition, rather than its defining features. I develop a diagnostic cure in Chapter2. The particular cure I develop is also antidotal and non-question-begging. It disarms arguments for skepticism about the external world using only observations which the skeptic can accept from within his skepticism. I accomplish this by showing the skeptic that such arguments depend on an unsupported intuition— namely, that perception can never guarantee that things are as we seem to perceive them to be. My diagnosis thus seeks to show that this intuition is the root cause of the relevant skepticism. If I am right, we may then come to view the skeptic’s denial of our knowledge to be a symptom of this root cause, rather than the defining feature of skepticism. Coming to view skepticism this way has significance for the necessity, method, and viability of a cure. First, emphasizing a root cause creates room for us to see the tendency toward skepticism as more widespread than it may otherwise appear: It might not just be

21 those who say that we cannot know many things, nor even just those who judge that we cannot, who are in need of a cure. It might also be those who harbor a feeling which naturally leads us to conclude that we cannot have such knowledge. The suggestion that a feeling can stand behind skepticism, and that those harboring it are skeptics, might be striking to those who view skepticism as a theory adopted through rational reflection on our perception or knowledge, or as a paradox which such reflection can bring to light.27 I will come back to this suggestion in §2.6 and develop it in Chapters3 and4. For now, it is enough to notice that tracing the skeptic’s denial of our knowledge to a root cause creates some room for expanding our conception of who counts as a skeptic. And if more people are afflicted with skepticism, the production of a cure becomes more vital. Second, if the root cause of skepticism is an intuition, the management of skepticism may often benefit from the management of the feelings and temperaments which influence our intuitions.28 Recognizing this can expand our methods for treating skepticism. This is a point I think Hume’s discussions of skepticism get right. I discuss Hume’s conception and moderation of a skeptical temperament in Chapter 3. Finally, viewing skepticism more in terms of its root cause and less in terms of an explicit denial of knowledge can, I think, lead us to recognize that skeptics are less firmly entrenched in their skeptical outlook than proponents of the preventive response make them out to be. Some skeptics proudly avow their commitment to a denial of our knowledge. But even their commitment proves to be on shakier ground than it may have first appeared, if it turns out to be propped up on an unsupported intuition. Such a finding would be good news for the viability of cures for skepticism, since a less steady skepticism suggests greater openness to other perspectives, which may then be fairly exploited as part of a cure. This last point suggests that cures of a different sort than the one I offer in Chapter2 are also viable. My cure in that chapter avoids using premises which are inconsistent with and so unacceptable to the skeptic’s position. But if skepticism is by its nature shaky, even cures which do not avoid such resources may be able to the reach the skeptic. I discuss

27For the latter conception, see Wright(1991a); Byrne(2004); Pritchard(2015); Echeverri(2017). 28Even when skepticism is adopted through rational reflection, it may not follow that only rational reflection can figure in a treatment of skepticism. Those who view the maintaining of skepticism in the face of arguments against it as irrational should find it especially natural to address the skeptic’s feelings as part of a cure. Feelings are thus a natural next place to look, at least for those who are not epistemic permissivists, and perhaps for permissivists as well. On permissivism, see n. 11.

22 one such cure in Chapter4. This is Kant’s cure for skeptical empiricism. As I read him, Kant offers a cure to a skeptical empiricist, despite its deep inconsistency with the skeptic’s position. In a certain sense, then, Kant’s cure begs the question against the skeptic, and is unacceptable from the skeptic’s position. But Kant believes it can appeal nonetheless. He thinks an instability and dissatisfaction inherent to skepticism allows the skeptical empiricist to look beyond her own position and so entertain the cure which Kant offers. If Kant is right, even ‘question-begging’ cures can reach the skeptic. There are further methods yet for curing skepticism which I do not have room to address. I will not discuss any potential supplemental cures. Nor will I discuss attempts at antidotal cures which try to argue that the skeptic’s commitments are inconsistent or incoherent. It is worth noting that thinking such attempts must fail is one of the common motivations for opting for mere prevention.29 Nothing I will say here counters these claims. Nor do I show that these strategies must fail. Instead, my argument in this chapter motivates experimenting with many kinds of cure. And the broad contours of a successful diagnostic response which I sketched within the dilemma can inspire hope that those contours can be filled in and developed into a potent cure. I now turn to this task.

29See §2.2, especially note 16. For a more hopeful prognosis of this strategy, see Goldhaber(ms).

23 2.0 THE GROUNDLESSNESS OF SKEPTICISM

2.1 PLAN FOR A CURE

Perception puts us in a position to know many things about a world of physical objects. This is a fact. But it is easy to lose our sense that it is true. Arguments for skepticism about the external world, the view that we have no knowledge whatsoever of how things are in the world around us, are plentiful and not easily resisted.1 These arguments threaten our understanding of ourselves and the world we live in. An ambitious or curative response to skepticism attempts to change the skeptic’s mind using premises and inferences the skeptic might accept. Many philosophers now believe such a response is bound to fail. They opt for a more modest, preventive response, which attempts only “to show how to retain. . . our pretheoretical beliefs about perception. . . without accepting the premises the skeptic needs for his argument.”2 Since the preventive response makes no claim to convince the skeptic, it may exploit assumptions that beg the question against him.3 In Chapter1, I argued that the merely preventive response is likely to fail: Until we can show a skeptic how she has gone wrong, we cannot easily resist her arguments in good faith. In what follows here, I show that we need not settle for mere prevention. I oﬀer a three-part cure. In §2.2, I show that prominent arguments for external world skepticism rely on a

1I primarily have in mind skepticism in the contemporary sense: a view or theory that we know very little. See, for example, Nozick(1981), 197; Stroud(1984), vii. But I also treat it as a condition which a person can have or be in. This kind comes in many varieties, including: a belief that we know very little; doubt that we can have knowledge of a particular kind; doubt about the truth of a class of propositions; a feeling of perplexity, discomfort, or alienation when considering claims to knowledge; awareness of an inability to understand how one can know what one takes oneself to know; a disposition to occupy any of the previous conditions; and so on. See my discussion in the Preface. 2Pryor(2000), 517. 3For clear endorsements of the preventive response to skepticism over the curative one, see Pryor(2000), 517–20; Williamson(2000), 27; Byrne(2004), 300–03; Pryor(2004), 370; Byrne(2014), 278, 285; DeRose (2017), 69–69, 84–86. Compare the further citations in Chapter1, note 11. In what follows, I will speak of cure and mere prevention, rather than ambitious and modest responses to skepticism, for the reasons I discussed in §1.2. I have also questioned, in §1.5, how comfortably a merely preventive response to skepticism can exploit assumptions that beg the question against the skeptic, but I leave that aside here. On the other hand, I will argue in Chapter4 that such assumptions can play a role in cure as well.

24 single, shared presupposition, which I call No Guarantee. This is the idea that perception can never put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be. In §2.3, I show that the best available arguments for this idea are question-begging. It follows that, unless a better argument can be given, the idea should be treated as a baseless intuition. In §2.4, I sketch an alternative view on which perception can put us in touch with the world in a way that guarantees what we seem to perceive. It does so when we in fact perceive things to be as they are. This view is sometimes called ‘epistemic disjunctivism.’ Since the view offers a straightforward way of making sense of perceptual knowledge, acceptance of it cures skepticism. I argue that once the external world skeptic is made to see that her skepticism is groundless, she can accept this alternative view on ordinary grounds, and so be cured. I then argue, in §2.5, that my three-part cure is uniquely able to answer several of the most pressing objections typically raised against anti-skeptical strategies which appeal to epistemic disjunctivism. Doing so not only clarifies how my cure’s three parts fit together. It also demonstrates that the cure is more dialectically effective, and able to reach a wider range of skeptics, than other disjunctivist strategies.

2.2 A COMMON PREMISE IN SKEPTICAL ARGUMENTS

Skeptical arguments often invoke so-called ‘skeptical scenarios.’ If you are now hallucinating, or dreaming, or being misled by a computer program or an evil demon, everything could appear to you exactly as it would if you were actually perceiving how things are. Things could be radically diﬀerent than they appear without your being able to tell. There would be, as Descartes said, no “signs” by which you could tell wakeful perception from that which is dreamt or simulated.4 I think we must accept, along with the skeptic, that such scenarios are possible.5 Let

4Descartes(1641/1996), 13. More carefully: the meditator of Descartes’ ﬁrst meditation says this, while the meditator of the sixth meditation acknowledges a sort of coherence in waking life which dreams lack. 5At least in the metaphysical sense, that nothing about the basic structure of reality precludes their possibly obtaining. It does not follow from this that they must be epistemically possible—that we cannot know them not to obtain. I will later suggest that we can know that they do not obtain, if they do not. For worries about the metaphysical possibility of such scenarios, see Echeverri(2017).

25 us also concede that if we are in such a scenario, we cannot know anything about the world around us through perception. Nor could we know that we are in such a scenario. These concessions to the skeptic do not yet imply that we cannot know anything about the world around us. But they can seem to, with the help of one or more very intuitive premises.6 Take, for example, the skeptical argument from the principle that knowledge is closed under known implication. According to that principle, if you know something, and you know that what you know implies something further, then you know that further implication.7 The argument runs as follows:

Argument from knowledge closure (A1) If you know that you have hands, and you know your having hands implies that you are not a disembodied brain in a vat, then you know that you are not a disembodied brain in a vat. (A2) You know that your having hands implies that you are not a disembodied brain in a vat. (A3) You do not know that you are not a disembodied brain in a vat. So, (A4) You do not know that you have hands.

The premises imply the skeptical result (A4) deductively. We must deny one of the premises, if we are to resist the skeptical result. This may seem tough, if the premises appear plausible. (A1) is an instance of the knowledge closure principle, which gives a plausible explanation of how we can extend or adapt our knowledge base through deductive inference. In this regard, it carries a theoretical load. To deny it would be to deny what seems to be one of the primary ways we learn new things or retract our claims to knowledge.8 What about (A3)? Many philosophers view this premise as an intuitive reaction to

6See, for example, McGinn(1989), 7: “At no point does [a skeptical] argument require us to accept anything that is either obviously false or even open to doubt.” See, also, Greco(2007), 2: “skeptical arguments are powerful in the following sense: it is not easy to see where they go wrong, and rejecting them requires one to adopt substantive and controversial theses about the nature of knowledge and evidence.” Similar points are made in DeRose(1999), 1–2. For an opposing view, see Byrne(2004). 7We can express this formally as follows: If you know φ, and you know φ entails ψ, then you know ψ. There are many variants of this principle. For our purposes, an imprecise formulation will suﬃce. For discussion, see Williamson(2000), 117; Kvanvig(2006); Hawthorne(2013), 29; Pritchard(2015), 12–14. 8Very few philosophers deny even weak versions of knowledge closure. Well-known exceptions include Dretske(1970); Nozick(1981), 204–09.

26 skeptical scenarios. They typically claim that it can be accepted in virtue of its intuitiveness.9 But we can ask on what conditions it appears intuitive. In general, we can try to make plausible a denial of our knowing that some contingent state of affairs does not obtain as follows: Say a little bit about what it would take to know that it does not obtain, and observe that we do not have or have not done what it would take. Suppose I am trying to make it plausible that we do not know that the Queen of England does not keep a diary. I may say, first, that in order to know that she does not we would have to ask her, or ask someone close to her, or go digging about through her bedroom drawers, or something of the like. I may say that we have not done any of these things, since we live far away, and she is a busy and well-guarded lady. I may add that even if she or a member of her retinue says that she does not keep a diary, this could be a lie—after all, diaries can be very personal, and not everyone who keeps one admits to doing so. I may also add that she could store her diary outside of the bedroom. Even if none of the ways that I suggested we could come to know that the Queen does not keep a diary turn out to be decisive, ruling some of these out may still help make it plausible that we do not know that she does not. In short, I can give all sorts of specific considerations which make plausible why this particular thing is something we do not know—considerations that specifically have to do with the Queen and diaries. Officially, (A3) denies your knowledge of a specific, contingent state of affairs—that is, of your not being a disembodied brain in a vat. Can we make this denial of knowledge plausible by saying what it would take to know this does not obtain, and then observing that we do not have or have not done what it would take? I think the answer must be no. On the one hand, it may seem that to know we are not a disembodied brain in a vat, we must simply check to see if we have hands or legs, and are not suspended in liquid in a vat. If that is all it takes, we can easily make that step, and probably have already, and so it will not seem plausible that we do not know we are not brains in vats. On the other hand, it can seem that even looking to see if we have hands would not be enough to show that we are not brains in vats, and indeed that anything we could try to check could not possibly confirm that we are

9For example, Pritchard(2015), 163 calls (A3) a “deep-seated intuition.” See also Nozick(1981), 167; Cohen(1988), 91; DeRose(1995), 2, 27, 48.

27 not. If that is right, it seems there are no specific things that we have not done which we could appeal to that could make plausible that you do not know you are not a disembodied brain in a vat. No considerations specific to brains and vats seem to play a role. This suggests that whatever makes it plausible that you do not know you are a brain in a vat is something very general. That this is so can be confirmed by noting that the argument goes through equally well if we substitute in more or less any skeptical scenario in the place of your being a brain in a vat. Any appeal to the specific state of affairs of being a brain in a vat turns out to be mere window dressing, whose only purpose is to dramatize the general idea. That idea is one whose truth would make it impossible to rule out our being brains in vats by appealing to what we seem to perceive: Even if we seem to perceive that we have hands, the idea goes, this could not guarantee that we do. We might then express the general idea in terms of what perception can guarantee to be true about the world it seems to put us in touch with. This all suggests that premises like (A3) strike us as intuitive only because we presuppose that perception can never put us in touch with the world in a way which guarantees that things are as we seem to perceive them to be. Let us call this presupposition No Guarantee. Consider what would follow if No Guarantee were false. If you do in fact see, or feel, that you have a hand, then this would put you in touch with the world in a way that guarantees that you have a hand. Such a guarantee would put you in a position to know you have a hand. You could then know that you are not a disembodied brain, by deduction from your perceptual knowledge.10

10I will not attempt to offer an analysis here of the notion of a guarantee, for three reasons. First, non- philosophers who have read or talked with me about this chapter have found the notion intuitive, and, indeed, have returned to it to better understand theoretical notions such as that of conclusive grounds. This leads me to suspect that an analysis of ‘guarantee’ may mislead us about which related notions to take as basic. Second, many of the central notions of epistemology are contested and used in disparate ways, including the notion of grounds, or conclusive grounds. I myself have serious doubts about the notion of defeasibility, and would resist understanding guarantees in terms of indefeasible reasons or evidence. Although I hope to return to this in later work, a discussion of the relation of guarantees to these other notions would be long and complex, taking us fairly far afield at this point. Third, the central presupposition at issue is not that perception offers a ‘guarantee’ of a particular sense or kind, but that perception offers no guarantee, of any kind, that things are as we seem to perceive them to be. In considering the presupposition, it is therefore useful to remain somewhat neutral about the nature of the guarantee. Nevertheless, I think epistemologists hoping for more familiar theoretical notions can, if they prefer, understand ‘guarantees’ in terms of conclusive grounds, without much effect on the argument of this chapter. Later on, I will assume that guarantees at least provide conclusive grounds for knowledge, whether or not they simply are those grounds.

28 I have not yet denied No Guarantee. At this point, I only ask us to imagine denying it, in order to bring out that the alleged intuitiveness of (A3) rests on accepting it. Could (A3)’s intuitiveness derive from another source? As Keith DeRose notes, it can be tempting to explain our inability to know that skeptical scenarios do not obtain by appealing to our inability to rule such scenarios out.11 But we can ask, again, why we ﬁnd it plausible that we cannot rule out skeptical scenarios. And, again, acceptance of No Guarantee seems central. After all, if we were to deny No Guarantee, it would be natural to think that some cases of perception allow us to rule out skeptical scenarios, for instance, by deduction from the perceptual knowledge they make available. The same holds for a variety of related claims about our apparently poor epistemic position with regard to skeptical scenarios—for example, that we cannot tell or discern whether they obtain, or distinguish their obtaining from their not obtaining. As DeRose notes, these claims are so closely related to each other that “citing one of these to explain the plausibility of another doesn’t occasion even the slightest advance in our understanding.”12

11DeRose(1995), 16. 12DeRose(1995), 16. According to DeRose, what makes our epistemic position with regard to skeptical scenarios seem bad is our intuitively thinking that knowledge requires sensitivity to the truth. One way to spell out this thought is as follows: you do not know something if you would believe it even if it were false. DeRose and others call this principle Sensitivity (DeRose, 1995, 38). DeRose thinks we “instinctively” select skeptical scenarios such that our beliefs that they do not obtain are insensitive. Even if we were brains in vats, or dreaming, he thinks, we would believe that we are not (DeRose, 1995, 18, 37). Perhaps a commitment to Sensitivity would indeed make premises like (A3) seem plausible. But that alone would not show that a commitment to Sensitivity is often, or even ever, what underlies (A3)’s plausibility. To see this, we can first note that many who feel the attraction of (A3) do not hold Sensitivity at all. Some may think we can know we performed some arithmetic correctly, even if, had we made a mistake, we would still have believed we were correct. They could not think this, if they were committed to Sensitivity. Others hold that we do not believe anything at all in certain skeptical scenarios. They think that when we dream, for example, we typically do not believe, but only dream that we believe. They must then deny that we would still believe we were not dreaming, even if this belief were false. Once again, people who think this can still find it plausible that we do not know we are not dreaming. See, for example, Sosa(2007), 9ff. This plausibility is unlikely to derive from Sensitivity for those who deny Sensitivity. Moreover, Sensitivity is not the only way to capture the idea that knowledge must be sensitive to the truth. A consideration of No Guarantee suggests a different way. Suppose I know that there is a maple tree in front of me. It may be true that if I were a brain in a vat, I would still believe that there is a maple tree in front of me. But I would not believe it on grounds which guarantee that there is a maple tree in front of me. I would believe it on different grounds. This is another way in which my knowledge can be sensitive to the truth. The mere presence of belief is not the only dimension of sensitivity to the truth. My justification, or the grounds of my belief, can be sensitive as well. If this is right, knowledge may indeed require sensitivity to the truth, without Sensitivity being true. No one believes that belief is all there is to knowledge. It can then seem mysterious why the notion of sensitivity to the truth would be spelled out as Sensitivity. We can now ask: Why would someone believe Sensitivity in the first place? A natural answer seems to be that Sensitivity appears plausible, because it

29 Without an independent source of plausibility, we should thus view any skeptical argument which employs premises like (A3) as tacitly presupposing No Guarantee.13 In other words, premises which deny that we can know whether skeptical scenarios obtain rely on No Guarantee for their plausibility. No Guarantee is controversial, and far from obvious. Any skeptical argument which presupposes it would be incomplete without a defense of that presupposition. In the next section, I will argue that such a defense is hard to come by. But we should first ask if any skeptical argument can avoid presupposing No Guarantee. The most promising way to avoid this presupposition, it would seem, would be to avoid any premise like (A3)—any premise which by itself places skeptical scenarios beyond our ken. After all, (A3) was the premise whose apparent plausibility was found to presuppose No Guarantee. Let us then look at a skeptical argument which makes no assumption of this kind, and ask whether it, too, presupposes No Guarantee. I will focus on the skeptical argument formulated by Jim Pryor in “The Skeptic and the Dogmatist.”14 Pryor’s main aim in this article is to give a preventive response to skepticism. But in order to do this, he first formulates the most powerful skeptical argument he can. To do so, he explicitly avoids assuming potentially controversial claims like (A3).15 For this very reason, Pryor’s seems that, beyond the presence or absence of a belief, there is no other relevant way in which our relation to the truth can vary between the ordinary and the skeptical case. This plausibility is likely to depend on No Guarantee. Unless we believe that perception can never put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be, we are unlikely to spell out sensitivity as Sensitivity. We will be able to notice other important differences between the cases: for example, the presence of a guarantee. Although accepting Sensitivity may lead some people to accept premises like (A3), Sensitivity is not itself a natural way of articulating the ways in which knowledge is sensitive to the truth. Instead, a commitment to No Guarantee is likely to underlie a commitment to Sensitivity. It is then once again No Guarantee, rather than an independent thought about sensitivity to the truth, which makes premises like (A3) seem plausible. 13Some arguments for skepticism that assume a premise like (A3), but do not rely on any closure principle, include: Wright(1991a); Brueckner(1994); DeRose(1995), 1; Lewis(1996), 539; Wright(2008), 403; Mizrahi (2016), 370–71. A similar case can be made about skeptical arguments from other epistemic principles. According to the principle of underdetermination, if a given piece of evidence does not support one claim over a second, then we cannot know the first claim on the basis of that evidence alone. With this principle in mind, a skeptic might claim that perception never provides better evidential support for your having hands than for your being a disembodied brain in a vat stimulated to think you have hands. Since things can appear the same in waking perception and certain skeptical scenarios, we can say that appearances are common to both. But why think our evidence must depend on that? The answer, I think, is again prior acceptance of No Guarantee. If perception did put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be, then in certain cases perception could provide conclusive evidential support. No other evidential support is as good as conclusive evidential support. 14Pryor(2000), 524–29. 15He writes: “some philosophers refuse to allow the skeptic to use claims like ‘I can’t know I’m not being

30 skeptical argument appears more forceful than the argument from knowledge closure. But, as I will argue, this skeptical argument too appears plausible only on the presupposition of No Guarantee. Pryor’s skeptical argument begins with a disjunction that even those who reject (A3) should ﬁnd plausible:

(B1) Either you do not know that you are not a brain in a vat or you know it on the basis of knowledge gained through perception.

Most people who reject (A3) now do so on the grounds that perception may, at least in some cases, put us in a position to know that skeptical scenarios are false.16 If the ﬁrst disjunct holds, it is relatively clear why skepticism would follow; the argument from knowledge closure is one possible route.17 The second disjunct states that you know you are not a brain in a vat on the basis of knowledge gained through perception. The skeptic thinks this disjunct is false. She thinks you cannot know that you are not a brain in a vat on the basis of perceiving, for example, that you have hands. She might say: “How can you be so conﬁdent in what you seem to be perceiving? You may think you are perceiving that you have hands. But how can you be sure that you are really perceiving this, and not just seeming to perceive it, if you haven’t already ruled out that you are a brain in a vat being fed false experiences by a computer program?” To rule out the second disjunct, then, Pryor’s skeptic “require[s] us to know that we’re not being deceived. . . antecedently to knowing anything on the basis of perception.”18 Pryor

deceived’ as premises in his reasoning. Maybe skeptical argument can convince us that we can’t know we’re not being deceived; but why should we grant such a claim as a premise in a skeptical argument?” (Pryor, 2000, 522–23, 543n11). 16(B1) is very close to the claim that we cannot know a priori that we are not in skeptical scenarios, a claim many now accept as a lesson learned the hard way. Transcendental arguments like those found in Strawson(1959) argue a priori that things in the world must be a certain way if we are to think or experience as we do. But the anti-skeptical merits of these arguments have been questioned. See Stroud(1968/2000), 24; Strawson(1985), 22–23; Cassam(1987), 377–78; Grayling(2010); Stroud(1994/2000), 162–63; Stern (2000), 59–63. Content externalist approaches like that of Putnam(1981), Ch. 1 argue a priori from the meaningfulness of thought or language to the falsity or unintelligibility of claims about being in skeptical scenarios. But the prospects of this approach have seemed dim. See Brueckner and Ebbs(2012); Miracchi (2017), 369–71. Some hold that a priori generic presumptions of the reliability of perception are rational. See BonJour(1985). But their number is small. Anyone committed to one of these a priori strategies may deny (B1). But most philosophers now accept it. 17Pryor’s skeptic could alternatively appeal to (B2), discussed below. 18Pryor(2000), 524.

31 spells out “antecedence” in terms of what one’s reasons for belief, or basis for knowledge, “presuppose or rest on.”19 The demand is for us to establish that we are not in a skeptical scenario in a way that does not rest or depend on the thought that we are now perceiving, and on that basis know, how things are in the world around us. If we cannot do this, the skeptic says, our perceiving things to be a certain way cannot be an adequate basis for our knowing that we are not in a skeptical scenario.20 We can express this requirement as follows:

(B2) If you know that you have hands on the basis of some perceptions, then you must be able to know that you are not a brain in a vat on some independent basis.

The general idea behind (B2) is that, in order to adequately ground knowledge of how things are in the world around us, perceptions must be supplemented with independently based knowledge that we are not in a skeptical scenario. (B1) says that we can have no non-perceptual basis for knowing that we are not in a skeptical scenario. So the independent basis must be provided by a different set of perceptions. But (B2) suggests that no set of perceptions itself ever provides a sufficient basis for this knowledge. So the buck gets pushed back further to another set of perceptions, and then to another, and so on indefinitely. (B2) generalized then implies that the second disjunct of (B1), that you know you are not a brain in a vat on the basis of perception, is false. It then follows deductively that we cannot know anything about the world on the basis of perception.21 We can now ask whether, and on what conditions, the premises are plausible. I will not consider the credentials for (B1) here.22 Why think that (B2) is true? To explain why it is plausible, Pryor introduces the notion of skeptical “badness.” Something is a “‘bad’ alternative just in case it has the special features that characterize the skeptic’s scenarios—

19Pryor(2000), 525. A curious feature of Pryor’s gloss of “antecedence” is that the relation he describes is not asymmetric. If you know that this tree is a maple and know that lemons are sour, you can know these things on bases that do not presuppose each other. You would then know each of them “antecedently” to the other. This suggests that the notion at issue is a notion of independence rather than antecedence, or “epistemic priority” (524). In the body text, I will thus speak of independence rather than antecedence. 20Pryor(2000), 524. 21Above, I suggested that a knowledge closure principle like (A1) could bridge the gap between the ﬁrst disjunct of (B1) and the skeptical conclusion. But (B2) would work just as well. Thus, we need not include (A1) within the premises of Pryor’s argument. See note 17. 22Though see note 16.

32 whatever those features turn out to be.”23 For Pryor, the hypothesis that what looks like a zebra is actually a painted mule might be bad by being “incompatible with what you purport to know,” while still being compatible with the grounds you have. The hypothesis that you are dreaming, or deceived by an evil demon, might be bad by “introduc[ing] a non-standard explanation of your experiences,” even though that hypothesis is compatible with your actually having hands. In either case, the skeptical hypothesis is bad insofar as it has features that “undermine the support your experiences give you for your perceptual beliefs.”24 Pryor uses this notion to “add two new premises:” A “Skeptical Principle about Knowl- edge” and a premise about the relevance of skeptical scenarios to the bases for knowledge provided by perception. They are as follows:

SPK: If you’re to know a proposition p on the basis of certain experiences or grounds E, then for every q which is ‘bad’ relative to E and p, you have to be in a position to know q to be false in a non-question-begging way—i.e., you have to be in a position to know q to be false antecedently to knowing p on the basis of E. (6) The hypothesis that you’re being deceived by an evil demon is ‘bad’ relative to any course of experience E and perceptual belief p.25

Suppose that p is “I have hands” and E is the ground for perceptual belief given by my experience of seeing my hands. There may be some scenarios that undermine E. If E is undermined, I cannot know that I have hands on the basis E. SPK implies that if I am to know that I have hands, I must be in a position to rule out every scenario that undermines E. In other words, I must have some basis for knowing that I am not in a ‘bad’ scenario which undermines my perceptual grounds for believing that I have hands. That basis must be independent of E itself, since at this point E may be vulnerable to being undermined. (6) tells us that there is some scenario that is ‘bad’ with respect to all of our experiences and perceptual beliefs. In other words, the threat described by SPK is real. We thus have (B2).

23Pryor(2000), 527. 24Pryor(2000), 527. In Pryor’s terminology, a “‘bad’ alternative” may not be bad relative to the same “proposition,” “experience,” or “perceptual belief” (528) to which it is an alternative. My dreaming can be bad relative to my having hands. But if I have hands in the dream, the dream is not an alternative to my having hands. It is an alternative to my being awake and perceiving that I have hands. This terminology can be confusing, so in what follows I avoid talk of alternatives. 25Pryor(2000), 528. Here, I keep Pryor’s numbering, though elsewhere I have used my own, drawing on Pryor selectively. The relevant numbered claims in Pryor are (5)–(9) and SPK.

33 That is, there is a skeptical scenario we must be able to rule out, if we are to know that we have hands; and we need some basis for ruling out this scenario which does not depend on the perception that we have hands. Pryor thinks that SPK is the source of (B2)’s skeptical power, and the claim which the skeptic has “work to do. . . to persuade us of.”26 SPK is the claim he then tries to resist in order to avoid skepticism.27 Notably, Pryor does not even mention (6), when considering the merits of the premises in his skeptical argument.28 The suggestion is that (6) is highly plausible, even undeniable. Is it? In one way, (6) is trivially true. That premise tells us that there is at least one skeptical scenario that is “bad” relative to all of our perceptual grounds. To be bad is, once again, to have “the special features that characterize the skeptic’s scenarios—whatever those features turn out to be.” Paradigmatic skeptical scenarios will always have the characteristic features of skeptical scenarios. An evil demon or vat scenario has the characteristic features of being unsettling and reminiscent of science ﬁction. These scenarios also have the feature that, if one of them is true, we could know nothing on the basis of perception. These features are central to what makes each scenario paradigmatically skeptical. But the kind of badness at issue is more than this. It is a kind which makes skeptical scenarios relevant to SPK and leads to a skeptical conclusion. The question here is not about the narrative genre of

26Pryor(2000), 529. 27Pryor(2000), 531–32, 537. More precisely, Pryor’s primary target is a slightly stronger claim which describes an analogous relationship between perceptual justification and the basis for perceptual belief given by a perception. He calls this claim “SPJ” (531). Ultimately, though, Pryor thinks rejecting this claim suffices for rejecting the slightly weaker claim about knowledge, since the former implies the latter (cf. 521). The latter is my focus. Pryor makes no attempt to say what work the skeptic might have to do to defend SPK or SPJ. Nor does he try to say why the skeptic might find either plausible. Pryor’s lack of interest in the skeptic’s motivations reappears in his response to the skeptical argument. He tries to counter that argument by advancing “dogmatist theories of perceptual knowledge and justification.” According to such theories, “when you have an experience as of p’s being the case, you have a kind of justification for believing p that does not presuppose or rest on any other evidence or justification you may have” (Pryor, 2000, 532). In effect, then, the dogmatist theories deny that the skeptic has any reason to demand that you know that skeptical scenarios are false antecedently to your having all the justification you need in order to know something about the world around you. That is why Pryor thinks the theories have “anti-skeptical punch” (537). What goes missing is a diagnosis of why the skeptic seemed to have a reason for this demand in the first place. If Pryor’s dogmatist theories cannot speak to that, it is hard to see how they can help put skepticism to rest. Pryor’s only defense of his dogmatist theories—that they can be shown to be plausible according to the “standard philosophical methodology [of] sensible philosophical conservatism” (538)—then does not seem to help. I raised a similar, though more general, criticism in Chapter1. 28Pryor(2000), 529.

34 skeptical scenarios. It is about whether a skeptical scenario is ‘bad’ relative to all perceptual grounds in the sense of undermining them. When understood in this way, I think, (6) appears plausible only if one assumes No Guarantee. If perception can never guarantee that things are as we seem to perceive them to be, then it is very plausible that there could be no grounds given by a perception which would not be undermined by some skeptical scenario. But imagine that perception can provide us with a guarantee. It would thus provide us with conclusive grounds for knowing that things are as we seem to perceive them to be. Such grounds are not undermined by any skeptical scenario, insofar as they guarantee that things are as we seem to perceive them to be. The grounds can seem to be undermined, if we think we have the same grounds in the skeptical scenario as we do in the case of genuine perception. But, if perception can give guarantees, the grounds are not the same. The conclusive grounds are conclusive, whereas no perceptual grounds we have in a (relevant) skeptical scenario could be conclusive. Nothing about how things appear to you in such a scenario could guarantee that things are as they appear. It follows that, unless we assume No Guarantee, the scenario does not have the central feature that leads to skeptical conclusions. It is not ‘bad’ relative to all perceptual grounds in the sense of undermining them. If perception ever gives a guarantee, then there are some perceptual grounds that are not undermined by even the skeptical scenario that we are being deceived by an evil demon. Let us call these conclusive grounds E*. If we really were in that skeptical scenario, we would not have E*. But this leaves open the possibility that we have E* now. And if we do, the scenario is not ‘bad’ relative to our grounds. It does not undermine them as grounds for knowledge. (6) is then false. So, unless we accept No Guarantee, there is no compelling reason to accept (6). Moreover, it is only the idea that perception cannot give us conclusive grounds that makes SPK look central to skepticism. Recall that SPK says roughly the following: In order to know something on the basis of some perceptual grounds E, we need to be in a position to rule out on independent grounds each scenario which undermines E. When perception gives us conclusive grounds, there are no “bad” scenarios which undermine those grounds. So SPK is trivially satisﬁed, and raises no challenge. Pryor is then wrong to say that “everything

35 will turn on the principle SPK.”29 The truth of SPK is consistent with perceptual knowledge of how things are in the world around us. Pryor has thus misdiagnosed the skeptical power of (B2). Its true source is (6), or rather, No Guarantee, the idea which makes (6) seem plausible.

2.3 JUST AN INTUITION

In each of the arguments I have surveyed, a premise about what kind of knowledge, or bases for knowledge, can be made available through perception appears plausible only on the assumption of No Guarantee, the idea that perception never puts us in touch with the world in a way that guarantees that things are as we seem to perceive them to be. We should, I think, treat this assumption as a common, often suppressed premise in skeptical arguments. The persuasiveness of such arguments then depends on the assumption’s merits. The assumption makes a sweeping claim about one of our cognitive faculties. It is not obviously true, and we have seen that it leads quickly to skepticism. What can be said in its favor? Presumably, an argument for No Guarantee would need to start from reflections on perception. These reflections may make claims about what I will call our epistemic position: our evidence, warrant, or knowledge. Or they may make no such claims. They can instead make claims about the basic character of perception. To reach No Guarantee, these claims would then need to be combined with further claims that bridge these basic characteristics with their alleged epistemic significance. Let us consider the first route. This route begins with claims about our epistemic position in perception. To avoid circularity, such claims must not depend for their plausibility on the assumption of No Guarantee—nor on any skeptical premise which depends on No Guarantee for its plausibility. If my arguments in §2.2 are sound, the resources are fairly meager. But some claims about our epistemic position in perception are not controversial. No one denies

29Pryor(2000), 529. This is not to deny that SPK may contribute to some skeptical conclusions in other cases in which perception does not provide a guarantee. Nevertheless, it is not the driving force behind a more general argument for skepticism about the external world.

36 that things can, through perception, appear to you one way when they are in fact another. In such cases, everyone agrees, perception fails to put you in a position to know how things are in the world around you. And in such cases, everyone agrees, it can still seem to you as if you are seeing how things are. You may not, and in some cases cannot, know that you only seem to perceive how things are, but are instead misperceiving, hallucinating, or undergoing an illusion rather than actually perceiving how things are. The possibility of skeptical scenarios attests to this fact. These uncontested observations can seem to imply No Guarantee, as follows:

(C1) When you merely seem to perceive, perception does not guarantee that things are as you seem to perceive them to be. (C2) When you merely seem to perceive, you cannot always tell that you are not actually perceiving things as they in fact are. So, (C3), your epistemic position can be no better when you are genuinely perceiving than when you are merely seeming to perceive and cannot tell that you are not genuinely perceiving. So, (C4), perception can never guarantee that things are as we seem to perceive them to be.30

The inference to (C3) can seem good insofar as your inability to know you are merely seeming to perceive can seem to level the playing ﬁeld. It can seem to show that, for all you know, you are misperceiving (or hallucinating, etc.) right now. After all, if you are in fact misperceiving, you may not know that you are. This thought could assail you even if it happens that you are perceiving things as they are. And that seems to imply that you are no better oﬀ even when you are not in fact misperceiving. If that is right, it would imply that you always lack a guarantee that you are perceiving things as they truly are. Any seeming guarantee would be just that: a seeming guarantee. But the inference to (C3) is worth questioning. Your not being able to tell you are misperceiving when you are, and your receiving something less than a guarantee in that

30An argument of this shape is sometimes called the “Argument from Illusion.” See discussions in McDowell (1982/1998), 385ﬀ; McDowell(1995), 878ﬀ; Reynolds(2000); Pritchard(2008), 294; Mizrahi(2016); French and Walters(2016); McDowell(2018).

37 case, does not logically imply that you cannot have that guarantee when you are perceiving things as they in fact are. This inference may seem justified, if we deny that perception can ever give you a guarantee that you are perceiving things as they are. But that denial is what the alleged bit of reasoning is meant to reach. If we help ourselves to it in order to vindicate the inference, our reasoning becomes question-begging. But if we do not help ourselves to the denial, it is not clear why we should treat the inference as justified. Merely entertaining the idea that perception can give us genuine guarantees is enough to show that the inference needs further support. But to provide such support would just be to give a different argument for the same conclusion. The original argument cannot hold itself up unaided. How does the other route to No Guarantee fare? This route exploits premises which state basic characterizations of perception and premises which bridge the characteristics with No Guarantee as their supposed epistemic consequence. Here, it is hard to see why one would accept the conjunction of both premises unless one already tacitly accepts the conclusion they are meant to support. Consider, for example, an argument which begins by characterizing perception as presenting us with mere images. On this conception, when you perceive a tree, you are presented only with an image of a tree. But your being presented with that image never itself implies that there is a tree there. Some further facts about the causes of the image would be needed to establish that. I think a bridge principle which links this ‘mere-image’ conception of perception to No Guarantee as its epistemic consequence would be quite plausible. For it is hard to see how being presented with images whose presentation does not essentially depend on how things actually are in the world could provide you with a guarantee of how things are in the world. But if one accepts this plausible bridge premise, it is hard to see why one should accept the initial characterization of perception as presenting us only with images, unless one accepts No Guarantee.31 Perhaps one may accept the conception on the ground that perception must present us with the same things both when we are perceiving and when we are merely seeming to perceive, but are instead

31One possible explanation is that one has so far failed to comprehend any alternative view of perception. That is part of why I offer an alternative in §2.4 below. As I will explain in Chapter4, Kant viewed Hume’s skepticism about perceptual knowledge as reflecting Hume’s failure to comprehend the mind’s operation in our perceptual knowledge. I briefly introduce this skepticism and Hume’s view of perception as composed of ‘fleeting,’ imagistic impressions in §2.7.

38 misperceiving, hallucinating, or etc. But, again, if one accepts the bridge premise, it is hard to see why one should think that we become aware of the same things in both cases, unless one already accepts No Guarantee. Other arguments may invoke a less restricted initial characterization of perception, and rely more heavily on the bridge premise to reach a skeptical conclusion. Say we allow that perception can, at least sometimes, make us aware of features of the world. We may try to reach a skeptical result using a bridge premise that implies that perception fails to give us a guarantee even when it makes us aware of how things are in the world. But what could lead us to grant such a premise? It is hard to see what could, except an argument like (C1)–(C4) above. We have already seen that such reasoning begs the question. And this suggests that the current reasoning, too, begs the question. It, too, assumes what it purports to show. In all of the above arguments, it is hard to see how one could reach No Guarantee without tacitly assuming it. The best reasoning in its favor is question-begging. The prospects for finding independent support seem dim. And without such support, we should treat No Guarantee as the bald and baseless intuition it is. In calling No Guarantee a baseless intuition, I do not mean to suggest that it is blatantly and unmistakably false, that not assuming it is a simple and trivial matter, or that the fact that skeptical arguments have relied on it for their plausibility is a great embarrassment. Any of these suggestions would give the impression that skepticism can be cured easily, and that a great deal of epistemology has been much ado about nothing. But the cure is not easy. Even if unsupported, No Guarantee can have a firm hold on us. When it does, a great deal of skeptical arguments appear plausible, even undeniable. As long as No Guarantee’s support for them remains tacit, it can be hard to pinpoint where the arguments go awry. And the more irresistible the arguments seem, the more entrenched the groundless intuition may become. It then takes great effort to identify and dislodge No Guarantee, first, by revealing the seeming support it gives to skeptical arguments, and, second, by canvassing and rejecting the most plausible routes of argument in its favor. This is no easy task. But carrying it out reveals the intuition to be nothing more than it is—an intuition lacking a firm foundation.

39 2.4 REACHING THE SKEPTIC

Timothy Williamson writes that “philosophical treatments of skepticism. . . are better at prevention than at cure. If a refutation of skepticism is supposed to reason one out of the hole, then skepticism is irrefutable.”32 Pryor agrees that “the ambitious anti-skeptical project cannot succeed.”33 But if I am right so far, we need not give up on reaching the skeptic. We can reach her, and change her mind, using the considerations given above. For the skeptic, no less than we, can be made to see that the seeming plausibility of her arguments relies on tacit acceptance of No Guarantee. And she, no less than we, can be made to see that she has no sound reasoning to support that idea. This can have a sobering eﬀect, though she may as yet see no other way to think about perceptual knowledge. All she needs to be cured is a genuine alternative she can embrace. Once the skeptic sees that No Guarantee is groundless, there is no barrier to her seriously entertaining its negation. This is the thought that perception can, in some cases, put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be. In successful cases of perception, one becomes aware of how things are in the world. When one has a guarantee, the awareness one gains could not possibly be had were the world not the way one is perceiving it to be. One is, as it were, directly in touch with some features of the world, and thus in a position to know some facts about the world in virtue of such awareness.34 A vignette can help show that the idea is both intelligible and familiar. You arrive on time to a family reunion. A few familiar faces are there, but as expected, most are running late. Your brother eagerly greets you, and starts a conversation. After ten or so minutes he

32Williamson(2000), 27. 33Pryor(2000), 520. 34This is not to say that whenever perception provides you with a guarantee you cannot nonetheless be made to feel uncertain that you are perceiving things as they are. You may become agnostic, doubt, or even deny what you genuinely perceive, because you have been misled in this instance into mistrusting your current ability to perceive or your environment (e.g., by having been falsely told that you are now on LSD or that a window pane is a TV screen). Even in such cases, though, perception can make you aware of how things are in the world around you. You may then come to believe, and know, some facts about the world on the basis of this awareness as soon as you jettison whatever false beliefs cast doubt upon your currently perceiving things as they in fact are. The guarantee such perception provides remains available to you, and is restored as soon as you leave behind the obstacles to your recognizing it for what it is.

40 says, “By the way, have you talked to mom? She wants to ask you something.” “She’s here? How do you know? I haven’t seen her,” you respond. “I can see her right over there,” he answers, pointing. Sure enough, you see her, too. We can learn a lot from this scene about the ordinary conception of perceptual knowledge and its basis—a conception we employ in everyday life whenever we assert, ascribe, challenge, or defend our claims to perceptual knowledge. Your brother demonstrates that he knows your mother is there by appealing to the fact that he sees her. This answers your question about how he knows. And why should it not? His seeing her there makes him aware of a fact—namely, the fact that your mother is there. It is normal to treat his perception as guaranteeing that things are as he seems to perceive them to be. The idea that one can, but does not always, become aware through perception of how things are in the world, and that this awareness can be all one needs in order to know that that is how things are, is sometimes called ‘epistemic disjunctivism.’ The idea gets this title, because, according to it, any alleged case of perceiving is either like this or merely seems to be like this. It is either a perception of things as they really are, which provides an adequate basis for knowledge that things are that way, or it merely seems to be a perception, but in fact is a misperception, hallucination, or etc., and so does not oﬀer such a basis. The disjunction here emphasizes the fact that the bases are not common between a genuine perception and what merely seems to be a perception. As we saw, emphasizing this diﬀerence can help us resist a tempting slide into skepticism. In presenting this ordinary conception of perception and its relation to knowledge, and showing that it accords with our everyday practices of talking about knowledge, we have provided the skeptic with a plausible replacement for the conception expressed by No Guar- antee. The skeptic can then accept that she can acquire knowledge in the ordinary way: by taking in the world through perception. She would then cease to be a skeptic. She would be cured.35 35More sophisticated concerns with epistemic disjunctivism—for example, that it is inconsistent with the current science of perception—are still possible. For this charge, see Burge(2005), 1–40 and Burge (2011). For responses on behalf of disjunctivism, see McDowell(2011), McDowell(2013), and Goldhaber (forthcoming-b). Such concerns could create obstacles to the skeptic’s coming to grips with our perceptual knowledge, and thus to her being fully cured. The possible barrier to cure which such concerns could pose creates a powerful reason to address them.

41 It may seem that the skeptic would still dismiss the ordinary conception of perception. For the skeptic previously resisted our ordinary ways of treating perception as putting us in a position to know things about the world around us. Why should she change on this front? The answer, I think, is that she presented her resistance as based on rational considerations— namely, the ones articulated by her skeptical arguments. These considerations have been undermined by the discovery that those arguments rely on a groundless intuition. Now perhaps the skeptic may dig in her heels and hold fast to her resistance even after it is shown to be groundless. In that case, she would have refused the cure. But any skeptic who sincerely claims to be an enemy of dogmatism, or friend of reason, will know better. She will be sobered by a cure which reveals that reason is not on her side. Still, some anxiety may linger. Nothing I have said proves that we have any knowledge of the world around us. If, after all is said and done, we are victims of a skeptical scenario, we do not have knowledge of the world around us after all. Skepticism would then be true. That possibility, which we have not ruled out, would be tragic, I suppose. But this possibility, as I have argued, in no way threatens the idea that perception can, when the world cooperates, reveal to us how things really are.36 The way to see whether we have any knowledge of the world around us is to do just that—look and see. If, by seeing a book or a tree before you, you become aware of how things are in the world, you will have your answer. You will come to know something about the world, without needing any additional proof. Perhaps there really can be such a proof. Nothing I say here shows or even suggests there cannot be. But to expect that we would need such a proof in order to know anything about the world around us, or to put skepticism to rest, is just a version of No Guarantee. It is a version of the groundless skeptical intuition that we always need something more than what

36The fact that some cooperation from the world is required need not rattle our sense that knowledge is the knower’s achievement. If Myla runs a marathon, that is her achievement. But certain conditions in the world beyond her control need to obtain for her to achieve this. She could not have run the marathon, had there not been ground for her to push her legs oﬀ, and oxygen to fuel her metabolism and muscle movement. But the fact that these conditions had to obtain for her to run a marathon in no way rattles our sense that her running it was her achievement. Similarly, the fact that the world’s presenting you with an appearance of how things are is a condition on your being able to know that things are as they appear to be should not rattle our sense that your knowing is your achievement. It is in some sense lucky for you that such conditions obtain, when they do, but your knowledge need not thereby be seen as a result of mere luck.

42 perception can give.

2.5 OBJECTIONS AND REPLIES

The parts of the cure I have oﬀered stand in a relation of mutual support. The disjunctivist conception of perception provides a well grounded alternative to the view of perception that skeptics presuppose. At the same time, tracing a range of skeptical arguments to a common and groundless presupposition is crucial for resisting two prominent lines of argument often made against disjunctivist anti-skeptical strategies in general. Seeing how my cure can resolve these recurring problems for disjunctivism will help to defend both disjunctivism and the cure itself, while also clarifying how the parts of the cure go together. This will in turn help us see how my cure is distinctive. According to John McDowell:

it constitutes a response [to skepticism] if we can ﬁnd a way to insist that we can make sense of the idea of direct perceptual access to objective facts about the environment. That contradicts the claim that what perceptual experience yields, even in the best possible case, must be something less than having an environmental fact directly available to one. And without that thought, this scepticism loses its supposed basis and falls to the ground.37

McDowell’s response has left many unimpressed. One objection concerns its being one-size- fits-all. Why should undermining one claim count as an adequate response to a phenomenon with such very diverse sources? Crispin Wright complains that McDowell “tends to write of skepticism as if it were a ‘frame of mind,’ or a kind of rootless anxiety or preoccupation, when the truth—as manifested in the recent contemporary discussion—is that we have to deal with a number of specific, sharply formulable paradoxes, differing in detail in significant ways.”38 It can then seem na¨ıve to expect that a single move—the replacement of just one idea with another—could diagnose or dispense with skepticism about the external world as a whole. 37McDowell(2008/2009), 228. See also McDowell(1982/1998), 393–94; McDowell(1995), 888n19. 38Wright(2008), 395. See also Wright(2002), 341n13 on the treatment of skeptical arguments in McDowell (1994), 112–13: “The reader may be surprised by the crude—almost caricatured—conception of sceptical arguments which McDowell betrays there. In effect, they are treated as merely a sort of unreconstructed obsession with our fallibility.”

43 The cure I offered addresses this concern. Although McDowell does not offer an argument to show it, the seeming diversity of arguments for skepticism about the external world is at least largely superficial. Such arguments tend to rely on No Guarantee.39 When we see that so much brush grows from the same root, we can uproot the whole weed in one go. A second kind of discontent with McDowell’s strategy questions the grounds upon which he recommends the disjunctive conception of perception. Many claim that work is left undone until it is shown that the ordinary conception is correct and the skeptical one wrong. David Macarthur thinks McDowell “needs to do more than to show that a non-sceptical way of thinking is available. As we have seen, that leaves the sense that scepticism is also available as an equally reasonable option.”40 As Crispin Wright puts it: “The mere depiction of more comforting alternatives is not enough. . . A draw. . . [is] all the sceptic needs.” This sort of draw would seem to be the skeptic’s win, insofar as it would call for agnosticism about whether we have knowledge of the external world.41 What would it take for the ordinary conception to triumph? McDowell’s critics are rarely explicit about this. But their criticisms suggest that the conception must be defended in a way that the skeptic could be made to accept. Duncan Pritchard, for example, says: “Given the limited argumentative support that McDowell offers in favour of his view, it is fair to charge him with [the] dialectical vice” of “question-begging,” of “groundlessly assum[ing], in [his] premises, a claim that [his] dialectic opponent will not accept.”42

39McDowell’s official target is the “highest common factor” view of perceptual justification. This is the view that the epistemic position in a case of genuine perception can be no better than the epistemic position in the case of an indistinguishable misperception or hallucination. See McDowell(1982/1998), 386; McDowell (1994), 113; McDowell(2002), 99; McDowell(2008/2009), 231. Given that misperceptions and hallucinations do not give guarantees that things are as they appear, the highest common factor view straightforwardly implies No Guarantee. McDowell does not argue that skepticism in general depends on No Guarantee, though he arguably hints at such a view with his strong emphasis on skeptical arguments for claims that straightforwardly imply No Guarantee. See note 30 above. 40Macarthur(2003), 189. 41 Wright(1985), 444. In contrast, see Byrne(2014), 278: “The sceptic bears the onus of proof—if she can’t supply a reason why we do not in fact have ‘direct perceptual access,’ and if this is the most plausible way of explaining our perceptual knowledge, then we may fairly take ourselves to have such access.” For an intermediate perspective, see Millar(2008), 597, who thinks “we should not underestimate the amount of headway that McDowell makes against certain skeptical arguments,” yet still admits that “McDowell from time to time appears to be dismissive of skepticism. He sometimes gives the impression that once we reinstate a commonsense picture on which we are open to the world, directly embracing worldly facts, skepticism ceases to be of interest.” 42Pritchard(2009), 478. Pritchard continues: “McDowell himself never explicitly offers [sufficient] supporting argument [for the disjunctivist conception], and rests content instead to offer his view in a broadly

44 McDowell is, however, wary of letting the skeptic’s “tendentious ground rules” govern his discussion.43 To him, “the dreary history of epistemology” attests to the fact that once we accept the skeptic’s premises, “hope is rather faint” for rescuing a commonsense position.44 For this reason, McDowell’s official “response” to skepticism “is not to answer the skeptic’s challenges, but to diagnose their seeming urgency” as deriving from a particular source.45 Since the challenges need not be answered, he thinks: “my move is not well cast as an answer to skeptical challenges; it is more like a justification of a refusal to bother with them.”46 Not surprisingly, some have read this as “an official refusal to take scepticism seriously.”47 This refusal, suggested by McDowell’s remarks, underestimates the potential of a cure. McDowell himself shows some degree of diagnostic ambition. “My idea,” he says, “is that skepticism looks urgent only in the context of a visibly dubious assumption, which imposes a certain shape on the space of epistemological possibilities.”48 But his critics are right to think this idea needs a systematic defense. I offered one in §§2.2–2.3. Even more important is the point both McDowell and his critics miss: With such a defense in place, it is possible to give a compelling answer to the skeptic’s challenge without taking on her ‘tendentious ground rules.’ This is what the presentation of an alternative does, when accompanied by a convincing diagnosis. I don’t think this is just na¨ıve. Say that you have an upcoming appointment across town, and you learn there is a parade blocking traffic. You won’t be able to drive around the parade in time. And the appointment is too far for walking. So you conclude you cannot get to your appointment in time. You might, perhaps in a fit of exasperation, challenge me with the question, “How can I get to my appointment on time, if all my options will take too long?” If I respond that the subway goes beneath the parade, and a train is coming in five minutes, I have answered your question. But I have answered it, in part, by rejecting quietistic manner (as if simply outlining the main contours of the position would suffice for his audience to recognise its truth, and thereby exit the fly-bottle of scepticism). Perhaps there are some philosophical issues that are best approached in this manner, but scepticism is not one of them.” See Pritchard(2008), 302 for a related discussion. 43McDowell(2008/2009), 229. 44McDowell(1995), 878. 45McDowell(1995), 890. 46McDowell(1995), 888n19. 47Wright(2002), 341. 48McDowell(1995), 888n19.

45 the second clause—by explaining that there is an option you have not seen. I may need to say a little about why the subway really is an available option, if my answer is to reach you. But I do not need to accept what you prematurely concluded: that all your options will take too long. The availability of the subway contradicts this. Similarly, we can understand the skeptic’s challenge as follows: “How can I know anything about the world, if perception can never make me aware of how things are?” It counts as an answer to her question to explain that she has overlooked an available option: that perception directly puts us in touch with the world in some cases, though in others it only seems to. Of course, the skeptic has lost her sense that this is a live option. That is why she formulates her challenge as she does. But this is just where the diagnosis comes in. If the skeptic can be shown that her acceptance of No Guarantee is groundless, and that this groundless acceptance is the only barrier to her seeing another live option, the option can again become live, and she can be convinced to accept it. In this way, I think we can offer an answer to skeptical challenges. There is no need to be quiet when faced with them. We can talk to the skeptic. Such a conversation may well succeed, if it includes the right parts in the right order. Here, the disjunctivism McDowell emphasizes is the key final part, which is supported by and rightly follows the others. Showing the skeptic the groundlessness of his skepticism prepares him for the presentation of an alternative he can find compelling.

2.6 CONCLUSION

I have oﬀered a three-part cure for skepticism: (1) Trace skeptical arguments to a common presupposition, No Guarantee; (2) show that this presupposition is groundless; (3) oﬀer a viable alternative that does not lead to skepticism. This cure is meant to convince an external world skeptic to change her mind. It can defuse a variety of skeptical arguments. And although I have focused on external world skepticism, it can in principle be extended to other forms of skepticism. Those skeptical about other minds, or the reliability of memory, or the claims of morality, or meaning in life, can be reached in the same way, whenever their

46 skepticism can be traced to a similarly groundless presupposition, and replaced with a viable non-skeptical alternative. The three-part strategy, if executed successfully, provides all that is needed to change the relevant skeptic’s mind. The first two stages show her that her skepticism is the result of a baseless intuition, and that arguments which seem to imply it topple without prior acceptance of that intuition. This clears any obstacles to her coming to accept, in the third stage, an alternative to the intuition. The third stage answers her challenge to make sense of a kind of knowledge that she denied we have. One might still wonder to what extent this strategy is new or distinctive. Other anti- skeptical strategies may also question some common premise or presupposition of skeptical arguments, and try to show that its grounds are lacking or insufficient. They may well be accompanied by, or partly consist in, an alternative conception that is meant to avoid skepticism. Even if they do not explicitly take this form, they are likely to conflict with or challenge some premise or assumption in skeptical arguments, and to endorse or lead to some alternative conception. Their difference from the three-part strategy I describe can then seem to be superficial at best. For example, a contextualist response to skepticism may try to show that a variety of skeptical arguments appear plausible only on the assumption of an incorrect semantics for the word ‘know.’ The contextualist may then claim we can resist the skeptic’s conclusion by coming to appreciate that our standards for knowledge attribution depend on conversational context. I might be rightly called ‘tall’ in a kindergarten but not on a basketball court. Similarly, for a contextualist, I might be rightly said to know I am not a brain in a vat in at least most ordinary contexts, though in some special contexts I might not count as knowing it. At least in certain contexts, then, we might plausibly reject, for example, the premise that we do not know we are not a brain in a vat, or the premise that knowing that we are not is a prerequisite to knowing things about the world through perception.49 For the contextualist, skepticism might then be plausible only on the mistaken assumption that truths about knowledge are insensitive to context. In tracing the plausibility of some skeptical premises to a certain semantical theory, and questioning the grounds for that theory

49For contextualist rejections of the former premise, see Lewis(1996); DeRose(2017), 64–66.

47 in part by arguing for an alternative semantical theory, does the contextualist not give an instance of the cure I am proposing? In fact, I think there are several important differences. First, the basic shape of my anti-skeptical strategy differs from this one. I do not question the grounds for No Guarantee by arguing for an alternative conception of perception or knowledge. In my three-part cure, the second part comes before the third. I canvas and consider the best arguments for the presupposition and find them to be question-begging independently of any considerations in favor of an alternative conception. Coming to see the presupposition as groundless can then open the skeptic to considering and accepting an alternative. A contextualist who appeals to the context-sensitivity of truths about knowledge in order to undermine skeptical arguments is advancing a view which the skeptic may so far take himself to have good grounds to reject. The cure I offer loosens or dislodges the skeptic’s underlying intuitions before even attempting to propose an alternative. Second, apart from the basic shape of the strategy, the cure I offer is distinctive in identifying No Guarantee as a key presupposition of skeptical arguments. I argued that No Guarantee is presupposed by a wide variety of skeptical arguments, making the cure distinctive in its breadth. Moreover, whereas presuppositions about the context-insensitivity of truths about knowledge are not obviously related to skepticism, No Guarantee is itself particularly close to skepticism about the external world. It expresses in a very general way that perception alone can never give us what we would need in order to form knowledge on its basis. This makes skeptical arguments that presuppose No Guarantee at least close to question-begging. That in turn makes it easier to undermine the apparent grounds for this presupposition without depending on an anti-skeptical alternative that the skeptic is not yet ready to accept. In other words, my cure is distinctive in the presupposition it targets, which in turn makes a cure of its distinctive shape more likely to succeed. Third, the alternative I offer is an ordinary conception. It can be debated and defended, but it is deeply familiar to all of us, including skeptics, in our everyday attitudes toward our perception of houses, trees, and people. Nothing is more familiar than the idea that we can know the world and things in it by seeing, hearing, and feeling them. It is not news that your seeing your mother gives you a guarantee that she is there. If I am right, the skeptic

48 does not need new reasons to accept this. She can accept it on ordinary grounds that are already familiar and once again available. This allows the cure to be especially appealing, and to avoid depending on contested theoretical claims like those about the context-sensitivity of ‘know.’ Some have thought that establishing these claims would require a great deal of empirical research.50 But no independent inquiry must be made before my cure can be accepted, beyond reminding the skeptic of, and clarifying, the things we already and ordinarily believe. It is worth remembering here that different anti-skeptical strategies may not be in compe- tition with each other. For all I have said here, contextualism may be correct, and arguments in its favor may then undermine an often hidden presupposition of skeptical arguments. In- deed, avoiding appeals to contested theoretical claims may help to make the cure I offer especially compatible with a range of other cures, as long as they are compatible with the idea that perception can provide guarantees. My cure may even make other kinds of cure more palatable to the skeptic, by removing a presupposition that can stand in their way. Skeptics might then be more willing to accept, for example, any empirical grounds appealed to by contextualism. On the other hand, the ordinariness of my alternative, the near-question-begging character of appeals to No Guarantee, and the defusing of skeptical arguments before offering an alternative combine to make my cure especially dialectically effective. The more argument is required to put in place a theory such as contextualism, the more likely it is that arguments for the theory will invoke premises which skeptics will not accept. My strategy also has distinctive implications for the role of skepticism in theorizing about human knowledge. According to many, that role is to be a measure by which to test our theories and principles. If an epistemic theory or principle can be shown to lead to skepticism, it is thereby shown to need refinement.51 While this picture is not wrong, it is

50See Buckwalter(2017). 51See Miracchi(2017), 364: “Skeptical challenges can show us where our epistemological theories have gone wrong. If we find that our theorizing motivates premises that compellingly lead to skeptical conclusions, we must have gone wrong somewhere in our theorizing. The anti-skeptical task for the epistemologist is to figure out what went wrong—how her theorizing about knowledge and justification could motivate such problematic conclusions.” See also Gupta(2019), 28: “The skeptic. . . provides us with a useful touchstone: a model of experience and rationality that allows skeptical arguments to go through is for that very reason a flawed model.”

49 apt to mislead. Or, to put the point another way, the ‘measure’ provided by skepticism could easily be abused. When principles such as knowledge closure are made to lead to skepticism, it would be a mistake to fault the principles, and on that ground suggest all sorts of apparent refinements to them. The refinements would not help to avoid skepticism, so long as No Guarantee is held in place. All this would do is leave us with less plausible principles, now neutered or chimerical. The kind of ‘measure’ skepticism is, first and foremost, is a dowsing rod for finding No Guarantee, or an analogue, at the root of one’s theory. Relatedly, if I am right, one need not conclude from the emergence of skepticism that one is working with too demanding a notion of knowledge. If the skeptic demands that our knowledge be based on a guarantee, he is not asking too much. Instead of lowering the standards for knowledge, we can show that perception can rise to them. We can do this by showing that the barriers to seeing that it can are groundless. I want to close with a brief reflection on some implications that go beyond theoretical pursuits altogether. According to McDowell, we become skeptics as a result of faulty philosophy. It is “a familiar epistemologists’ syndrome”52 which causes us to lose sight of a conception of perceptual knowledge which “ought to seem sheer common sense, and it would if questionable philosophy did not put it at risk.”53 McDowell’s anti-skeptical project is “to rescue the position from bad philosophy, and to leave it looking perfectly satisfactory.”54 This gives the impression that the project exhausts itself in correcting a theoretical error. Faulty philosophy may indeed be the main source of the seemingly rational support structure which helps hold skepticism in place. But philosophy need not be the original source of skepticism itself. Skepticism regularly occurs to people who have no training in philosophy or rhetoric and little inclination to speculate, including even children. Everyone “gets” The Matrix. This may just seem to show that philosophical ideas are pervasive in our culture, and more accessible than academic philosophers often treat them to be. After all, what is the The Matrix but a pop, sci-fi twist on Descartes’ dreams and demons? But I doubt that this is the whole story. Indeed, I find it very plausible that the source of

52McDowell(1995), 878. 53McDowell(2002), 98. 54McDowell(2002), 98.

50 skepticism need not lie in any kind of reasoning. Since No Guarantee enters our minds as a groundless intuition—as is strongly suggested by the fact that its best available defenses are question-begging—it can take hold of us through feeling. Anxiety, depression, feelings of powerlessness, excessive brooding, and especially alienation can lead us to systematically underestimate what we ourselves, including our perception, are capable of. Such feelings can untether us from the world, and make even the most familiar things feel far away, sometimes too far away for even our perception to reach. A sustained bout of such feelings, even if only subdued and in the background, can lead us to believe No Guarantee. Even when philosophy causes one to adopt No Guarantee, such feelings can also hold that intuition, and so skepticism, in place. Consider a skeptic who follows the cure I offer—that is, he comes to see that skeptical arguments rely on No Guarantee, that this presupposition is a groundless intuition, and that a familiar, non-skeptical alternative conception of perception is available. If he still continues to deny that we can know things about the world around us, it may be his psychology rather than his philosophy which now needs treatment. His skepticism may be influenced more by unsettling feelings than by rational reflection. I think even those who treat skepticism as a view adopted as a result of rational reflection, or as a paradox which can emerge within it, should nonetheless acknowledge that feelings may stand in the way of its resolution.55 In fact, they should acknowledge another kind of connection between skepticism and feeling as well. Skepticism, even when adopted as the result of some philosophy, can in turn leave troubling feelings in its wake. It can make the world, and other people in it, feel unreal. It can alienate us from our environment, and each other, and make things seem to matter less than they do. If I am right, such feelings may often be the original sources of our skepticism. But they can also be its consequences, when skepticism combines with, magnifies, and seems to justify preexisting isolation, anxiety, apathy, egocentrism, or sociopathy. In this way, the value of the cure I offer here is not limited to correcting a mistake in philosophical theorizing about knowledge. It does not just promise to “cause a sea of philosophy to subside.”56 It

55This case may be aptly described by Wittgenstein(2005), §86: “Difficulty of Philosophy not the In- tellectual Difficulty of the Sciences, but the Difficulty of a Change of Attitude [Schwierigkeit einer Umstel- lung]. . . Philosophy does require a resignation, but one of feeling [des Gefühls], not of intellect. And maybe that is what makes it so difficult for many.” 56McDowell(1982/1998), 389.

51 contributes to a defense of human cooperation, commerce, psychology, and morality.57 The simple idea that we can really be in touch with the world is a powerful one. Removing the barriers between us and that idea, and so between us and the world, helps us reclaim a sense of belonging to a world we share with others.

2.7 TRANSITION: HISTORICAL PRECEDENT

I have argued that skepticism about the external world, one of the guises of skepticism about perceptual knowledge, rests on an unsupported intuition: No Guarantee. This finding is in principle available to the skeptic. Communicating the finding along with an alternative conception of perception and its relation to knowledge constitutes a cure for his skepticism. We need not give up on developing a cure, as many philosophers now do (§1.3). Nor must we ourselves remain perplexed by skeptical arguments (§1.5). Instead, we can reach the skeptic, satisfying our humanitarian aim (§1.5). In Chapters3 and4, I will shift my focus from contemporary epistemology to historical discussions of skepticism from the eighteenth century—in particular, Hume’s and Kant’s.58 My interests in what these figures have to say on the matter are many. First, I will look to their discussions to find a historical precedent for the conception and treatment of skepticism I have put forward in the Preface and Chapters1 and2. In particular I will find that both philosophers, in their own way, find certain moods and temperaments central to skepticism. This resonates not only with my proposal in the Preface to view skepticism as more than a view about human knowledge, including also frustration or alienation felt with regard to our knowledge. It also resonates with my finding in §2.3 that skeptical arguments rest on a bald intuition—one which I suggest in §2.6 can arise from and be held in place by feeling. Moreover, I find in Kant a viable cure for skepticism, and in particular one which follows similar contours to the one I have just offered. The first step of Kant’s cure, as of mine, is to

57For a classic treatment of altruism as based in a conception of oneself as one person among others who are equally “real,” see Nagel(1978). 58For a note on abbreviations, editions, and citation styles used for Hume and Kant’s writings, see my “Note on Citations of Hume and Kant” below, beginning on p. 160.

52 show that the kind of skepticism at issue—one which he takes Hume to have championed— is based on insufficient grounds. Also like my cure, Kant’s terminates in an explanation of our knowledge which the skeptic has so far failed to comprehend. I discuss Kant’s cure in Chapter4, and the ways it clarifies, justifies, and explains the efficacy of my own in Chapter 5. Second, I think both Hume and Kant’s discussions draw attention to what we can learn from talking to the skeptic, apart from curing her. If we not only speak but listen, we can learn how to moderate our own tendencies to dogmatism and overzealousness. In Chapter 3, I argue that Hume viewed skepticism as apt for this task because he conceived of it as a temperament opposed to and capable of curbing a confident or sanguine temperament. In Chapter4, I explain that Kant viewed skepticism as a welcome wake up call, showing the need for a proper inquiry into the limits of human knowledge. If these philosophers are right, skepticism plays a crucial role in keeping our belief from soaring to excessive heights. Third, I have chosen to discuss these philosophers within the context of this dissertation in part because I think aspects of my discussion so far can prime us to better understand these philosophers on their own terms. In particular, my proposal to view the treatment of skepticism as the treatment of a person opens new and easily overlooked ways to read Hume and Kant. It does this, first, by allowing us to view skepticism as a condition a human being can have or be in. Distancing skepticism from a view about human knowledge can create room for understanding skepticism as Hume saw it: as a temperament tending towards melancholic excess. I argue that we should reconceive Hume’s skepticism along these lines in §3.5. My proposed view on the treatment of skepticism helps us read the historical philosophers, second, by removing an overly constrained and overly combative portrayal of the dialectic with skepticism typical of contemporary epistemology. That portrayal has, I argue, obscured how Kant thought he could reach the skeptical empiricists whom he hoped to cure. It has, I think, led both to undue criticisms of Kant’s attempts at cure and to a misplaced denial that he meant to cure skeptics in the first place. I discuss both of these interpretive camps in §4.2 and replace them with my reading of Kant’s cure in §4.6. Fourth, it is plausible that the spread of a conception of perception often associated with Hume’s empiricism helps to explain, though not necessarily justify, why many people

53 easily adopt No Guarantee. This is a conception of perception as composed of very short- lived sense impressions. Many empiricists have tried to explain how successions of these fleeting sensations enable all of our various and sometimes complex thoughts about the world. Hume was engaged in such a constructive project. But he found that working with such resources required viewing much of what we take ourselves to know about the world through perception to be based on a “fiction,” “false supposition,” or “gross illusion” of the imagination, including, for example, the “continu’d existence” of objects while we are not perceiving them.59 He said: “Since all impressions are internal and perishing existences, and appear as such, the notion of [objects’] distinct and continu’d existence must arise from a concurrence of some of [our impressions’] qualities with the qualities of the imagination.”60 Hume thus seems to acknowledge that the sense impressions on which he grounds his theory of the mind can only acquaint us with something ‘thinner’ than the robust, lasting objects which we ordinarily take ourselves to perceive. From this he concludes that the beliefs we form in subsisting objects, standing in causal relations to one another, must be a result of the imagination’s fictions. Reflection on this predicament leads Hume to feel “more inclin’d to repose no faith at all in [his] senses, or rather imagination, than to place in it. . . an implicit confidence.”61 Indeed, if such beliefs can amount to knowledge at all, they do not seem to be knowledge of the objects we ordinarily take ourselves to know through perception. A conclusion of this sort is naturally seen as a kind of skepticism about perceptual knowledge; since it arises amidst an empiricist program, we can call it ‘skeptical empiricism.’62 It may

59T 1.4.2.36, SBN 205; 1.4.2.56, SBN 217; 1.4.2.56, SBN 217. 60T 1.4.2.15, SBN 194; cf. 1.4.2.20, SBN 195; EHU 12.6–11, SBN 151–53. 61T 1.4.2.56, SBN 217. 62I do not claim that what we can properly call ‘skeptical empiricism’ is exhausted by this kind of skepticism about perceptual knowledge. Historical empiricists are also known for a skepticism which they cast on our pretensions to know things through reason. Hume’s empiricist project, for example, leads to a “scepticism with regard to reason” (T 1.4.1.) in addition to his “scepticism with regard to the senses” (T 1.4.2). And, as we will see from Kant’s discussion of empiricism in Chapter4, a skepticism about the use of reason in metaphysics may be not only an eﬀect of empiricism, but also its cause. In Chapter3, I put aside the question of whether it is correct to attribute a kind of skepticism about perceptual knowledge to Hume. I do this largely so I can keep the focus on what Hume thought skepticism was. But I also wish to stay neutral on what turns out to be a surprisingly complex interpretive matter. Nonetheless, readers of Hume have found his arguments concerning the origin of our idea of the continued existence of objects or of our idea of necessary connection to lead to skepticism about perceptual knowledge, whether or not Hume thought they did. See Stroud(1977), 115–17; Fogelin(1993), 105–08; Strawson (2002); Ainslie(2015), Ch. 4. And others, arguably Kant included, have found a broadly empiricist account of perception as involving only ﬂeeting sense impressions to lead to skepticism about, or an inability to make

54 come as no surprise that holding the skeptical empiricist’s conception of perception makes perception seem unable to deliver guarantees. In this way, skeptical empiricism can make the external world skeptic’s arguments seem irresistible. The connection between skepticism and empiricist conceptions of perception helps motivate a turn to these historical figures, both of whom, I argue, offer resources for responding to this kind of skepticism. Fifth, I think the strategies outlined in Chapters2,3, and4 are most effective when they are combined. Hume offers a way to characterize and treat the characteristic moods of skeptics in a way that supplements my finding of the groundlessness of No Guarantee. And as I read Kant, he offers the empiricist a remedy for her predicament by explaining how certain concepts necessarily figure in our experience. These include the concept of ‘substance’—that is, of something permanent and unchanging in the object—and the concept of ‘cause,’ which situates objects in a continuous nexus of interaction with other objects. Presumably, if the skeptic can accept and grasp Kant’s explanation, she could come to see how perception can put us in touch with more robust objects than Hume’s quickly “perishing” impressions could. That suggests that Kant’s more detailed conception of perception can be combined with the one I offer in §2.4 to help skeptics better understand how perception can guarantee the truth of what we perceive. That could help skeptical empiricists, for example, out of their skepticism about perceptual knowledge. As I will suggest in Chapter5, my cure and Kant’s can combine to make a more potent cure which reaches a wider range of skeptics.

sense of, perceptual knowledge. They have come to this conclusion independently of Hume’s arguments seemingly to that eﬀect. See Austin(1962/1964); McDowell(1994), Ch. 1; Stroud(2006), 343–45. There are certainly skeptics in the vicinity, and they can properly be called ‘skeptical empiricists.’ Some of them have been inﬂuenced by Hume’s works directly. And these skeptics are worth acknowledging and trying to cure, whether or not Hume was one, or thought himself to be one.

55 3.0 THE HUMORS IN HUME’S SKEPTICISM

3.1 AN OVERLOOKED QUESTION

In the striking concluding section (1.4.7) of the first book of A Treatise of Human Nature, David Hume pauses his project of “explain[ing] the principles of human nature”1 to survey his findings so far. Doing so fills him with “desponding reflections” about “the wretched condition, weakness, and disorder” of human cognitive faculties. Since Hume must use these faculties in his project, “melancholy” and “despair” replace his normal vigor for it, and “discourage [him] from further enquiries.”2 By the end of the short section, however, Hume finds himself again in an “easy disposition” of “good humour.” He feels ready to “contribute a little to the advancement of knowledge”—to bring his science of man “a little more into fashion.” He invites his readers to “follow [him] in [his] further speculations” in the Treatise’s latter books.3 Interpreters have placed great weight on this curious transformation. Understanding how Hume emerges from melancholy to again embrace philosophy has seemed to be the key to Hume’s considered views in the Treatise on the roles of skepticism and philosophy in human life.4 And so it has seemed crucial to understand the transformation’s structure and details. Any complete account of the transformation must note that it is far from immediate. Hume does not pass directly from his skeptical melancholy to resumed interest in philosophy. Rather, his transformation involves several distinct stages: After plunging into melancholy (1.4.7.1–8), Hume stops engaging in philosophy to enjoy social pleasures (1.4.7.9); then feels aggression toward philosophy (1.4.7.10); then composes his mind through restful activities (1.4.7.12); and finally finds himself curious and ready to resume his philosophy (1.4.7.12 and 14). The succession can be pictured as follows:

1For a note on abbreviations, editions, and citation styles used for Hume’s writings, see my “Note on Citations of Hume and Kant” below, beginning on p. 160. 2T 1.4.7.1–3, SBN 264–65. 3T 1.4.7.14, SBN 273. 4See Stroud(1991/2011), 144–45; Baier(1991), 1ﬀ; Livingston(1998), 27; Garrett(2002), 206–08; Harris (2015), 101; Ainslie(2015), 5, 246.

56 Hume’s stages in 1.4.7 melancholy → sociability → aggression → composure → curiosity

Why does Hume pass through these several stages, in this order, and no others? Inter- preters often miss the question, despite their interest in Hume’s transformation. Or rather, many are content to attribute Hume’s passage through these stages to ‘nature’ without further explanation. Barry Stroud, for example, stresses “the naturalness and virtual in- evitability of reflecting philosophically.”5 Hume does report at the end of the process that he is “naturally inclin’d” to reflect and “cannot forbear having a curiosity” about philosophical topics.6 But noting the naturalness of the process does not itself shed light on the significance or order of the stages that compose it. It does not explain why Hume’s specific progression, as opposed to any other, is the natural one. Annette Baier distinguishes several “swings in moods” along the way, noting that the “transitions” Hume undergoes “are motivated. . . merely by the incompleteness of the initial mood, its natural fate of supplementation by a successor mood.”7 But she does not specify in what sense the moods are incomplete or bound to lead to some specific successor. Similarly, Don Garrett views Hume’s transformation as involving several shifts of what he calls “moods” or, after Hume, “bents of mind.”8 But he does not explain why the aggressive mood is, as he puts it, a “natural successor to philosophical melancholy and delirium.”9 Nor does he explain how it is a natural precursor to renewed interest in philosophy. Some interpreters emphasize the role of two passions in Hume’s return to philosophy. When it comes to Hume’s decision “whether he should recommit himself to philosophy,” Henry Allison writes: “Fortunately for both Hume and us this decision is made for him by the re-emergence of the inclination to philosophize, fostered by the passions of curiosity and ambition.”10 On Karl Schafer’s view, “the real foundation of Hume’s rejection of radical skepticism and of his positive epistemology more generally” is “a distinctively Humean ac-

5Stroud(1991/2011), 155. 6T 1.4.7.12, SBN 270–71. 7Baier(1991), 20–21. 8See, esp., Garrett(2002), 231–37. 9Garrett(2008), 165. 10Allison(2008a), 324.

57 count of epistemic virtue” based in the satisfaction of passions like curiosity.11 According to Donald Ainslie, Hume’s “return to philosophy is driven primarily by his feelings.”12 But none of these interpreters offers an account of why these passions arise so prominently when they do. We seem forced again to point to nature, without understanding its operations. I think we can say more. First, we can expose and clarify an interpretive question that has not received due attention. This is the question of why Hume undergoes the exact succession of stages he does—why he moves from melancholy to sociability, to aggression, to composure, and finally to curiosity. Second, we can answer the question. We can explain the significance, order, and completeness of the stages. In doing so, we uncover a striking and unappreciated view of skepticism and its role in human life. One important clue is the distinctive language Hume uses when describing the succession of stages. There, he speaks of “melancholy”13 and “spleen,”14 and makes frequent allusions to the other “humours”15 of ancient and medieval medicine. In what follows, I argue that Hume’s repeated invocation of the four humors is the key to understanding why he portrays himself as moving through the stages he does. For Hume views his recovery from melancholy as, or as mediated by, a series of shifts in the temperaments of humoral theory. I call this ‘the humoral reading’ of 1.4.7, and develop it in §3.2 below.16

11Schafer(2014), 4. 12Ainslie(2015), 225. Ainslie divides 1.4.7 into five parts, but they do not match the five stages of Hume’s transformation that I have listed. He groups the second and third stages (‘sociability’ and ‘aggression’) under the label “splenetic sentiments,” and then groups the fourth and fifth stages (‘composure’ and ‘curiosity’) under the relatively uninformative label “the slow return of an inclination to philosophize” (222). I think it is worth distinguishing the stages: in the second, Hume ignores philosophy, while in the third he is revulsed by it. And in the fourth he composes his mind, something he must complete before the inclination to philosophize returns in the fifth. 13T 1.4.7.1, SBN 264; 1.4.7.9, SBN 269. 14T 1.4.7.11, SBN 270. See, also, “splenetic” at 1.4.7.10, SBN 269. 15T 1.4.7.10, SBN 269; 1.4.7.11, SBN 270; 1.4.7.14, SBN 273. 16Very few interpreters draw attention to the humoral terms, with most passing them over entirely. Allison (2008a) 319–26, for example, gives a detailed sentence-by-sentence analysis of Hume’s “spleen and indolence” paragraph (1.4.7.9), without any discussion of its mentions of ‘spleen’ or ‘good humour.’ Similarly, Fogelin (2009), 6–7, helpfully distinguishes “four contrasting Humes, or at least four contrasting voices of Hume, inhabiting” 1.4.7, without clearly likening them to the four humors. A notable exception is Broughton(2005), 189–90 who suggests in passing that “perhaps ‘humors’ are the best term” for describing several of the stages of Hume’s transformation. But she does not draw out the implications of the humoral language. Nor does she list all the stages or describe any in detail. Ainslie (2015) goes a step further, interpreting Hume’s appeal to humoral melancholy as meant to reveal the way in which too much engagement in philosophy can be damaging. He says: “My suggestion is that Hume deploys the rhetoric of melancholia in [1.4.7] in order to draw on the idea that it is a condition brought about by excessive study’s interfering with the body’s proper functioning. For I think that he recognizes that there

58 I then consider two natural objections to the reading. §3.3 addresses what I call ‘the anachronism objection.’ This is the charge that it is anachronistic to claim that Hume drew inspiration from an ancient theory of medicine. In response, I argue that humoral theory not only appears throughout Hume’s own writings and early sources, but also retained a modest influence over early eighteenth-century Scottish medicine. §3.4 addresses what I call ‘the methodology objection.’ This is the charge that the humoral reading portrays Hume as violating an apparent methodological commitment to avoid appealing to the bodily causes of mental phenomena. In response, I explain that Hume makes frequent appeals to human physiology. But I suggest that his appeals are meant to offer models for thinking about the mind, without thereby committing him to any one theory of the parts or functions of the body. If that is right, Hume’s invocations of the humors throughout 1.4.7 suggest that he models his progression on the four temperaments of humoral psychology, without his needing to endorse the associated physiology. In §3.5, I turn to the humoral reading’s most important philosophical upshots. The reading, I argue, sheds a new light on the nature and significance of skepticism. On that reading, 1.4.7 offers a picture of health or proper functioning as a balance of basic temperaments of mind. The humoral reading, together with other aspects of Hume’s discussions of skepticism, suggests that, for Hume, skepticism is primarily a temperament—a temperament which, when balanced with others, produces the proper doxastic disposition. This conception, I argue, helps to resolve an important puzzle about Hume’s view of skepticism. The puzzle concerns how Hume can consider skepticism to be both an illness and a cure. Conceiving of skepticism as a temperament resolves the puzzle, because a temperament can be an illness when overly predominant, and a cure when moderate. This resolution in turn presents additional evidence that Hume held such a conception.

is something about philosophy itself that leads it, in particular, to interfere with the proper functioning of the mind” (14). Ainslie only mentions one humor, however, and, relatedly, ignores the role of the others in reestablishing proper functioning. As I argue in the text below, Hume conceives of not only the problem but also the solution in humoral terms. Wicker(2016), esp. 54–57, too, gives a rich reading of 1.4.7 as employing tropes from medical and cultural discussions of melancholy, but overlooks the role of counterbalance through opposing humors or temperaments in Hume’s recovery. Some interpreters invoke the theory of humors and temperaments in connection with Hume’s letter to an anonymous physician (HL 3) without relating the theory to Hume’s philosophical works. See Watkins(2018), 11. Others have stressed the inﬂuence of physiology and psychology on Hume’s discussions of skepticism and melancholy in the Treatise without mentioning the humors. See Wright(1983) and Ievers(2015).

59 Though this conception of skepticism is shaped by humoral theory, it can interest us even if the theory is false. We do not need to conceive of health as consisting in a proper proportion of humors in order to appreciate Hume’s idea that skepticism is a temperament which, according to its degree of predominance, can both threaten and restore the mind’s proper functioning. This is a unique and illuminating conception of the nature, source, and proper treatment of skepticism. On that conception, skepticism is not a theory as much as a temperament; its familiar and threatening manifestations do not arise from argument as much as from temperamental excess; and their proper resolution comes less through counterargument than through counterbalance by other temperaments.

3.2 A CYCLE THROUGH THE HUMORS

According to humoral theory, health requires a balance or proper proportion of four basic bodily liquids, called ‘humors.’ Disease, both mental and physical, occurs when one of these humors is overabundant or deﬁcient, either throughout the body or in one of its parts.17 The theory’s four humors, canonized in the Hippocratic text Nature of Man, are blood, yellow bile (or choler), black bile (or melancholy), and phlegm.18 Hippocrates associated each humor with a season and stage of life in which it was thought to be abundant.19 Over the centuries, the humors took on further associations. Galen of Pergamon, for example, emphasized that each humor, like each season, was either hot or cold and either wet or dry, and was thus associated with one of the four elements of nature: air, ﬁre, water, and earth.20 Later, each humor became associated with one of four temperaments—roughly, clusters of psychological and physiological traits and dispositions. Each temperament involved a

17For canonical expressions of these tenets, see Hippocrates, Nature of Man, Ch. 4 in Jones(1931), Vol. 4, 10–13; pseudo-Galen, On the Humours in Grant(2000), 16. For short introductions to humoral theory and its role in aetiology, see Jackson(1969), 368–69; Porter(1987), 39–41. 18See Nature of Man Ch. 4–5 in Jones(1931), Vol. 4, 10–15. These four are not the only humors acknowledged in the Hippocratic corpus, let alone the whole of Greek medicine. They became canonical thanks to continued attention to Nature of Man by later medical writers, especially Galen. See Jones(1931), Vol. 1, xlix–l and Jouanna(2012), 335–38 for discussion. 19Airs Waters Places Ch. 10–11, 13 in Jones(1931), Vol. 1, 98–105 and 108–11; Nature of Man Ch. 7, 15 in Vol. 4, 18–23 and 38–41; Humours Ch. 12–14 in Vol. 4, 82–89. 20See the discussion of Galen’s On the Doctrine of Hippocrates and Plato in Jouanna(2012), 338–40.

60 tendency to feel characteristic passions, engage in certain actions, and develop physical features and ailments associated with related body archetypes. Though individuals, and whole nations, were often thought to be born with prevailing temperaments, a person’s temperament could sometimes change with shifts in life-stage, environment, or season. It was thought that the more of a given humor one had, the more its corresponding temperament was expressed, and vice versa.21 Humoral therapies aimed at restoring humoral balance by dissolving or curbing the production of excessive humors, or promoting the production or retention of deficient ones. This was accomplished first and foremost by the adoption of regimens for the so-called “six non-naturals”: food, evacuation of wastes, exercise, air, sleep, and passions. Activities like study, music, and various social pleasures were prescribed for the effects they had on one’s passionate, and so humoral, constitution. If the ailment was severe, more invasive treatment, like purgative herbs or blood-letting, was used to expel excessive humors. But this harsher treatment was generally reserved for cases in which diet and regimen would not suffice.22 The rich set of humoral associations provided a framework for disease prognosis, diagnosis, treatment, and prevention. Because blood was thought to predominate in the spring and in youth, for example, special precautions would be taken in those times to avoid and treat ailments of excessive blood, like hemorrhaging. Sanguine youths, for example, were made to avoid meats and other ‘blood-rich’ foods, vigorous exercise in the heat, and excessive mirth.23 One disease, perhaps more than any other, garnered perennial interest among writers

21The idea that regional climates and mores were responsible not only for regional illnesses and physiques, but also characters, exists already in Hippocratic writings. See, for example, Airs Waters Places Ch. 12–24 in Jones(1931), Vol. 1, 104–37. The association of humors with character types can be found already in Galen’s later works. See the discussion in Jouanna(2012), 340. But a more robust theory of four temperaments only emerged in the several centuries following Galen. See, especially, the pseudo-Galenic On the Humours in Grant(2000), 17. See, also, several texts in Greek and Latin that are either modeled on that pseudo-Galenic text or share a common model with that text, displayed in Jouanna(2012), 341–58. 22See the discussion in Robert Burton’s 1621 Anatomy of Melancholy 2.4.1.1, EL 210–12, which frequently cites Galen, among many others. This English-language text, which collects quotations from two millennia of medical, philosophical, religious, and literary treatments of melancholy and medicine, gives an impressively accurate and complete representation of the humoral tradition up until Burton’s own day. I thus cite it as an invaluable source. Citations to this text, hereafter abbreviated Anatomy, refer to the part, section, member, and subsection numbers, followed by page numbers from the 1964 Everyman’s Library edition (Burton, 1621/1964). Volume numbers correspond with Burton’s part numbers. 23See Hippocrates, Regimen in Health, Ch. 1, 7 in Jones(1931), Vol. 4, 44–47 and 54–57.

61 hot cold blood phlegm spring winter wet infancy (or all of youth) old age air (or equal mixture) water sanguine temperament phlegmatic temperament = social, hopeful, headstrong = lazy, forgetful, content yellow bile (choler) black bile (melancholy) summer autumn dry adolescence maturity ﬁre earth bilious (choleric) temperament melancholic temperament = aggressive, impulsive, bold = gloomy, careful, studious

Figure 1: The four humors

62 in the humoral tradition. This was melancholy: a lingering agitation, depression, or madness, without fever, that was usually attended with baseless fear and sadness.24 According to humoral theory, melancholy was the extreme expression of a dominating humor—in this case, black bile. Indeed, the word ‘melancholy’ comes from the Greek for black (melas) and bile (kholos). Cold, dry black bile was associated with reflection, pensiveness, idleness, and caution in moderate quantities, and depression, doubt, paralysis, and madness in excessive quantities. Black bile encouraged arduous, focused thought—the kind employed in mathematics, philosophy, and other sober or scholarly pursuits.25 And those pursuits encouraged more black bile, either directly26 or because they tended to inspire melancholic sentiments when they, through their content or difficulty, revealed our own cognitive limitations.27 Humoral allusions pervade Hume’s discussion of melancholy. Hume calls his melancholy a “deplorable condition,”28 and depicts it in a way that fits the classic humoral description of the ailment: a sustained depression with bouts of madness attended by unprompted fear and sadness. He describes his melancholy as a “delirium,”29 and himself as “affrighted” and “forlorn.”30 When Hume describes his survey of his philosophy’s skeptical strands as “heat[ing his] brain,”31 he alludes to one of the primary ways accumulated black bile was thought to cause melancholic delirium: Black bile in the abdomen or blood produces hot vapors which rise to the brain, heat it, and obscure thought.32 Accordingly, Hume describes

24See Anatomy 1.1.3.1, EL 169–70. 25This provides an answer to the following question from the Aristotelian/Theophrastian Problems: “Why is it that all those who have become eminent in philosophy or politics or poetry or the arts are clearly of an atrabilious [i.e., melancholic] temperament, and some of them to such an extent as to be aﬀected by diseases caused by black bile?” (Book 30, Ch. 1, 953a10–14, translation from Barnes, 1984). See also Anatomy 1.2.2.6, EL 247; 1.3.1.4, EL 406–08; 2.2.3.8, EL 206–07. 26Fixed or repetitive thinking and mental application were thought to dry the brain, exhaust bodily heat, and stall digestive processes, so that the food is converted into black bile in place of blood (see Anatomy 1.2.2.6, EL 245–49 and 1.2.3.15, EL 301–05). 27See, for example, the tenth-century Islamic physician Ishaq ibn Imram, who says: “all those. . . who overexert themselves in reading philosophical books, or books on medicine and logic, or books which permit a view of all things. . . assimilate melancholy. . . in the consciousness of their intellectual weakness, and in their distress thereat” (Klibansky et al., 1964/2019, 84–85). 28T 1.4.7.8, SBN 269. 29T 1.4.7.9, SBN 269. 30T 1.4.7.2, SBN 264. 31T 1.4.7.8, SBN 268. 32See Anatomy 1.1.3.1–4, EL 169–77; 1.3.3, EL 419–22. See also Timothie Bright’s 1586 A Treatise of Melancholly, 2: “Abundance or immoderate hotenesse. . . yeeldeth up to the braine certaine vapors, whereby the understanding is obscured.”

63 himself as tormented with “clouds”33 and a “storm.”34 Melancholics were thought to retreat to dark, solitary places and imagine themselves to have transformed into fearful beasts.35 In the throes of melancholy, Hume reports: “I fancy myself some strange uncouth monster, who not being able to mingle and unite in society, has been expell’d from all human commerce, and left utterly abandon’d and disconsolate. . . . inviron’d with the deepest darkness.”36 This leaves Hume craving, among other things, the “warmth of the crowd”37 to counterbalance black bile’s frigidity and tendency toward isolation. Hume does in fact break free from his melancholy by joining the crowd. He dines, plays backgammon, and makes merry with friends.38 In enjoying social pleasures, he partakes in the most common humoral remedies for melancholy. According to humoral theory, melancholy tends to nurture itself by encouraging strained and obsessive thoughts. Social pleasures combat this vicious cycle, by distracting us from these thoughts and promoting blood.39 Hume emphasizes these two points. He tells us that a “person of a sorrowful and melancholy disposition is very credulous of every thing, that nourishes his prevailing passion.”40 And, in the Treatise’s second book, he explains that our natural aversion to melancholy urges us to seek out the excitements of social life:

Those, who take a pleasure in declaiming against human nature, have observ’d, that man is altogether insuﬃcient to support himself; and that when you loosen all the holds, which he has of external objects, he immediately drops down into the deepest melancholy and

33T 1.4.7.9, SBN 269. 34T 1.4.7.2, SBN 264. 35See French physician AndréDu Laurens’s 1594 A Discourse of the Preservation of Sight: of Melancho- like Diseases; of Rheumes, and of Old Age, 82: “The melancholike man. . . maketh himselfe a terrour unto himselfe, as the beast which looketh himself in a glasse. . . [H]e can not live with companie. To conclude, hee is become a savadge creature, haunting the shadowed places, suspicious, soliterie, enemie to the Sunne, and one whom nothing can please, but onely discontent.” See also the following stanza of poetry which prefaces Burton’s Anatomy and summarizes melancholy’s psychological symptoms: “’Tis my sole plague to be alone, / I am a beast, a monster grown, / I will no light nor company, / I will find it now my misery. / The scene is turn’d, my joys are gone, / fear, discontent, and sorrows come. / All my grief to this are folly, / Naught so fierce as melancholy” (Anatomy, “The Author’s Abstract of Melancholy,” Vol. 1, p. 12). For a discussion of melancholy and lycanthropy, see Jackson(1990), Ch. 14. 36T 1.4.7.2, SBN 264; 1.4.7.8, SBN 269. 37T 1.4.7.2, SBN 264. 38T 1.4.7.9, SBN 269. 39See Anatomy 2.2.4, EL 69–99; 2.2.6.1–4, EL 109–26. See also Ishaq ibn Imran’s prescription of “pleasant discourse” (Klibansky et al., 1964/2019, 85). The use of purgatives and blood-letting was comparatively rare. See Anatomy 2.4.2, EL 225–34; 2.5.1.3, EL 238–41; 2.5.1.2, EL 237. 40T 1.3.10.4, SBN 120. Similarly, Hume says of melancholy that “’tis usual for that passion, above all others, to indulge itself” (T 1.4.7.1, SBN 263–64).

64 despair. From this, say they, proceeds that continual search after amusement in gaming, in hunting, in business; by which we endeavour to forget ourselves, and excite our spirits from the languid state, into which they fall, when not sustain’d by some brisk and lively emotion. To this method of thinking I so far agree, that I own the mind to be insuﬃcient, of itself, to its own entertainment, and that it naturally seeks after foreign objects, which may produce a lively sensation, and agitate the spirits. On the appearance of such an object it awakes, as it were, from a dream: The blood ﬂows with a new tide: The heart is elevated: And the whole man acquires a vigour, which he cannot command in his solitary and calm moments.41

Here, Hume agrees with the misanthropes that our distaste for melancholy drives a “continual,” and so inevitable, “search for amusement.” Such amusements, he agrees, “produce a lively sensation,” dispelling the melancholy. Notably, Hume adds a further, physiological explanation: The amusement dispels the melancholy by causing a “new tide” of blood. While this remark can naturally evoke the mechanistic physiologies that emerged in the late seventeenth century,42 it is equally at home in a humoral framework. In the latter, the connection to psychology is more direct. Warm, wet blood was thought to be the most abundant humor, and for that reason the most balanced and typically conducive to health. Accordingly, the sanguine temperament, brought about by a predominance of blood, was social, carefree, active, and optimistic, though sometimes headstrong. Hume’s use of the word ‘sanguine’ would seem to agree. For he speaks of “sanguine tempers” that are “social” and “sincere” but “impetuous,”43 as well as the “sanguine hopes of young adventurers.”44 It is thus plausible that he would see an increase in blood or circulation as helping to present philosophy in an attractive light—as an enterprise in which one may make real accomplishments. That could foster adventurous passions like curiosity and ambition. Hume’s merriments over dinner and backgammon play a pivotal role in his recovery. The infusion of warm blood they provide raises him from the darkest depths. It cuts short his obsessive, despondent reﬂections and invigorates him, breaking his paralysis. But Hume’s recovery is not yet complete. Philosophical speculations do not yet interest him, but rather appear “so cold” in contrast to the invigoration of social life that Hume “cannot ﬁnd it

41T 2.2.4.4, SBN 352–53. 42See the discussion of mechanistic physiologies in §§3.3–3.4 below. 43H 3.24.31; 3.27.11; 3.27.20. 44H 5.41.43. See, also, H 3.25.27; EHU 1.12, SBN 12.

65 in [his] heart to enter into them any farther.”45 The contrast is perspicuous because the blood infusion does not fully vanquish Hume’s melancholy, leaving significant “remains of [that] former disposition.”46 The extent of this melancholic residue may reflect the severity of Hume’s initial humoral imbalance. While melancholic, he “fanc[ies himself] in the most deplorable condition possible, inviron’d with the deepest darkness, and utterly depriv’d of the use of every member and faculty.”47 Correcting that degree of excess calls for drastic medicine. When regimen does not suffice, a purge may be needed to restore balance. Hume does not mention purgative medicines. But the next stage of his recovery is characterized by a desire for a purge of sorts. For Hume, the poison which must be expelled is philosophy. He reports, “I am ready to throw all my books and papers into the fire.”48 Of the four humors, yellow bile, or choler, was seen as most suited for purgation. Its heat and dryness were thought to encourage diarrhea and vomiting—hence, the name of the disease cholera. Likewise, many melancholy-purging drugs had hot and dry natures.49 This allowed them to dissolve and evacuate black bile which persisted through treatment by regimen. It is fitting, then, that Hume wishes to destroy his books with hot, dry fire—the element associated with yellow bile. Hume describes himself in this stage as “governed” by a “splenetic humour.”50 One of the senses of the word ‘splenetic’ is ‘melancholic.’ So it is tempting to think Hume is returning to black bile here—after all, black bile is thought to reside in the spleen, and some melancholic diseases to result from spleen dysfunction.51 But ‘splenetic’ has another, quite different meaning of ‘given or liable to fits of angry impatience or irritability; ill- humored, testy, irascible.’ The Oxford English Dictionary lists this sense as particularly common throughout the eighteenth century, when Hume wrote.52 And it seems to be the sense Hume has in mind, since the splenetic humor makes him impulsive and aggressive.

45T 1.4.7.9, SBN 269. 46T 1.4.7.10, SBN 269. 47T 1.4.7.8, SBN 269. 48T 1.4.7.10, SBN 269. 49Such as asarum, laurel, and sea onion. See Anatomy 2.4.2, EL 225–26. 50T 1.4.7.10, SBN 269. 51See Jouanna(2012), 231n4 and Jackson(1990), 9–10. 52Relatedly, in the seventeenth and eighteenth centuries, ‘to spleen someone’ meant “to have a grudge at” him or her (OED, S: 637–41). The entries on ‘spleen’ and cognates in Samuel Johnson’s 1755 Dictionary of the English Language agree.

66 These dispositions are more characteristic of the bilious or choleric temperament than the melancholic.53 And the presence of fire imagery further suggests that it is a surge of yellow bile which clears Hume of melancholic residue. Still, no curiosity or ambition emerges at this stage. For the yellow bile fosters a felt aggression toward philosophy, which now seems an enemy—“against the current of nature,” even “torture.”54 Hume’s curiosity and ambition return only after he has “tir’d with amusement and company, and [has] indulg’d a revery in [his] chamber, or a solitary walk by a riverside.” Only then does he report: “I feel my mind all collected within itself, and am naturally inclin’d to carry my mind” to philosophical topics.55 These quiet, calming activities and resulting composure are plausibly associated with the fourth and final humor: phlegm.56 People with phlegmatic temperaments are generally peaceful, inactive, and equanimous, when not just slothful or indolent. And Hume’s “riverside” walk gives a subtle nod to phlegm’s association with water.57 Hume’s phlegmatic activities are a fitting rejoinder to his previous splenetic humor, given that cool, wet phlegm is most opposed to hot, dry yellow bile. But they would have offered no therapy, and so would not have promoted balance, while Hume was under the influence of excessive black bile. At that stage, black bile would have filled those quiet and solitary moments with obsessive and despondent thoughts. But the same activities can be therapeutic now that blood has dissolved, and yellow bile expunged, the excessive black bile. With the added phlegm they provide, Hume can “collect” his mind. And once it is collected, he can emerge from the phlegm to engage in more focused thought. Having made a full cycle through all four humors, Hume finds himself in humoral balance. Health is restored, and with it the inclination to indulge in some reflection. Only now does Hume feel the “curiosity to be acquainted with the principles of moral good and evil, the nature and foundation of government, and the cause of those several passions and

53Anatomy 1.3.1.3, EL 401. 54T 1.4.7.10, SBN 269–70. 55T 1.4.7.12, SBN 270. 56Hume sometimes uses ‘phlegm’ to mean coolness or indifference. See E Sc 55, Mil 180. 57This association is reflected in the use of the word ‘phlegmatick’ in Hume’s time to describe the watery parts of solutions. Johnson(1756) lists “watry” as a sense of the adjective, and cites Newton’s usage in distinguishing the flammable part of wine from the “phlegmatick” part.

67 inclinations, which actuate and govern [him].”58 As if to drive home the requirement of balanced humors for curiosity, Hume invites readers “of the same easy disposition” as him to join him in further speculation, while advising those who are not to “wait the returns of application and good humour.”59 We can summarize the humoral reading of the transformation as follows:

Hume’s stages in 1.4.7, with corresponding humors melancholy → sociability → aggression → composure → curiosity black bile → blood → yellow bile → phlegm → health

Earlier, I asked why Hume undergoes the particular succession of stages in 1.4.7. The humoral reading answers this interpretive question. It explains why there are five stages, in the order they appear, with the characteristic sentiments and activities they have. The melancholic humor and temperament urges an eventual invigoration of the blood through social activity; this leaves significant melancholic remains, which must be purged through the aggression of yellow bile; the restoration of phlegm must come last among the individual humors, because so long as significant melancholic excess remains, phlegmatic activities will fill Hume with desponding thoughts and plunge him back into melancholy; and health comes after that because it requires a balance of all the humors. If the humoral vocabulary and imagery throughout 1.4.7 offered nothing more than a dramatic tone, then the interpretive question would remain unaddressed. But in fact Hume’s language can clue us in to the framework in which he conceives of his recovery—a framework which explains the succession of its stages. We can still ask: To what extent is this progression particular to Hume? Is Hume’s description of it meant to be more than a self-report? Is it more? If Hume’s imbalance is especially severe, his cure too might be somewhat unusual. It might not always be necessary, for example, to cycle through all four humors in order to reach a balance. Perhaps sometimes a new tide of blood from “some avocation” may suffice as medicine. Hume may perhaps even allow that a melancholic “bent of mind” could, in some cases, “relax” on its own.60 Humoral

58T 1.4.7.12, SBN 270–71. 59T 1.4.7.14, SBN 273. 60T 1.4.7.9, SBN 269. Compare T 1.4.2.57, SBN 218, where Hume describes a certain “sceptical doubt” as a “malady” that cannot be “radically cur’d” but whose proper “remedy” nonetheless is “carelessness and

68 theory does not always require the precise sequence of 1.4.7 to restore balance. Instead, the rich and systematic humoral language of 1.4.7 show vividly that Hume conceives of recovery in humoral terms.

3.3 THE ANACHRONISM OBJECTION

The humoral reading portrays Hume as drawing systematically from an ancient theory of medicine. But Hume is in many ways a modern thinker. He wrote in a time of “enlightenment,” often characterized in terms of its replacement of classical scientific theories with remarkable innovations. Reading Hume as appealing to an antiquated, and to us occult, theory of medicine can then seem anachronistic and fantastic. I call this ‘the anachronism objection.’ If the objection is correct, we might then read Hume’s uses of ‘melancholy’ to be like ours, referring to a sentiment divorced from any associated humor or temperament.61 Indeed, we might wonder whether Hume was familiar with humoral theory at all. A first step toward answering this objection is to show that Hume was indeed familiar with humoral theory. If the abundant humoral language of 1.4.7 does not convince us, there is strong evidence elsewhere in Hume’s corpus. In Dialogues Concerning Natural Religion, for example, Hume’s character Philo expresses a humoral conception of health: “On the mixture and secretion of the humours and juices of the body depend the health and prosperity of the animal,” as humors “frequently become pernicious, by [their] excess or defect.”62 Soon after, Philo speaks of the “600 different muscles” and “284 bones” of Galenic anatomy,63 suggesting the author’s knowledge of Galenic doctrines. In the Enquiry Concerning Human Understanding, Hume describes a painful case of gout as caused by “malignant humours in [the] body.”64 In the History of England, Hume describes no less than three historical figures

in-attention.” Inattention is, for Hume, necessary to lead us away from certain skeptical reasonings, and so out of the doubt or gloom which they can produce. But inattention alone may not always succeed in restoring our interest in philosophy. That is often more involved, as it is in 1.4.7. 61Hume does call melancholy a “passion” at T 1.4.7.1, SBN 264 and T 2.1.11.2, SBN 317. 62DNR 11.11, KS 209–10. 63DNR 12.3, KS 215. 64EHU 8.34, SBN 101.

69 as dying from melancholy.65 And in the Treatise, he seems to endorse the theory when he says, for example, that “the mixture of humours and the composition of minute parts may justly be presum’d to be somewhat different in men from what it is in mere animals.”66 In these passages, Hume suggests that our bodies are composed of humors and that humoral imbalance causes disorders. These are the fundamentals of humoral theory. Hume’s familiarity with humoral theory is also suggested by its presence in the philosophical and critical texts Hume read while preparing the Treatise. Cicero’s Tusculan Dis- putations, for instance, draws an analogy between the “disorder” of a mind plagued by false education and “distempers and sickness. . . bred in the body from the corruption of blood, and the too great abundance of phlegm and bile.”67 More notable yet are the vivid ref- erences to the theory of temperaments in AbbéJ.B. Dubos’s 1719 Critical reflections on poetry, painting and music, which Hume no doubt read in preparation for his planned portion of the Treatise on “Criticism.”68 These appear, for example, in Dubos’s explanations of how painters “bring us acquainted with. . . the temperament” of their subjects, using phys- iognomy, hair color, and posture.69 He writes of Charles-Antoine Coypel’s Susanna Accused of Adultery:

The painter has diversiﬁed the complexions of the famous old men [crowded around the accused]; one appears fresh and sanguine, and the other choleric and melancholy. The latter, pursuant to the proper character of his temperament, which is obstinacy, commits the crime with heat and resolution. Rage and fury spread through his whole countenance. The sanguine old fellow seems to relent, and, notwithstanding all the transport of his passion, feels a remorse that staggers his resolution. This is the natural character of men of that kind of complexion.70

In this passage and others, Dubos invokes groupings of imagistic, physiognomic, and psychological tropes that have their root in Galenic writings—the choleric’s face betrays the “heat” of his rage and fury. Such passages surely contributed to Hume’s ﬂuency with the

65The Earl of Marre at H 5.40.64; Queen Elizabeth at H 5.44.60; and Mrs. Claypole at H 7.61.89. 66T 2.1.12.2, SBN 325. 67Book IV, Ch. X, p. 136 in Cicero(1877). Hume says in “My Own Life” that he “devoured” Cicero’s writings in his early years (E MOL 3, Mil xxxii–xxxiii). In an early letter to Michael Ramsey, he speciﬁcally mentions “a Tusculan Dispute of Cicero’s” (HL 1.10). 68Promised at T Intro.5, SBN xv–xvi. 69Dubos(1748), 78. The ﬁrst French edition was printed in 1719. 70Dubos(1748), 82–83; see also 214–15 for comparison of the choleric-sanguine with the melancholic in Julio Romano’s “great alter” at St. Stephen’s.

70 psychology and imagery associated with the four temperaments.71 But even had Hume never read Dubos, these associations would still have been familiar from the literary canon. To deny this would be to claim that Hume never encountered or appreciated Shakespeare’s “life. . . made of four,”72 or Milton’s “melancholy damp of a cold and dry,”73 or any of Ben Jonson’s numerous “comedies of humours,”74 not to mention the frequent humoral allusions throughout Molière’splays.75 Educated eighteenth century readers—Hume included—were well acquainted with the use of humoral tropes to capture certain well-worn character types and to explore pivotal facets of human nature—for example, our caprice, tendency towards extremity, incompleteness in romantic longing, or mortality.76 In such a context, there was nothing bizarre about Hume’s drawing inspiration from humoral theory, whose imagery he rightly expected his readers to recognize. If Hume had detailed knowledge of humoral theory, it is not unnatural to take the abundant humoral language of 1.4.7 at face value. Hume would have understood what he was so repeatedly alluding to. Nevertheless, those familiar with Hume’s scientific context may still doubt that Hume meant these allusions as more than dramatic flourishes. Despite its enduring presence in European culture and thought, humoral theory was in decline, and, at the academies of the young Hume’s Scotland, had largely been replaced by more modern medical theories. There, like much of the early eighteenth century British Isles, a mechanistic physiology was in vogue.77 Contemporary Scottish mechanists like George Cheyne viewed

71Hume has two notes on Dubos in his extant memoranda from his years writing the Treatise. See Mossner (1948), 500, No. 2–3 in Section II (labeled “Philosophy”). Dubos’s Critical Reflections also appears in the library Hume passed on to his cousin at his death (Norton and Norton, 1996, 88, No. 385). 72Sonnet XLV. 73Paradise Lost, Part II, Book XI, lines 542–45. 74See, especially, Asper’s memorable discussion of the literal and figurative uses of the word ‘humour’ in the prologue to Every Man Out of His Humour. This begins with a brief capitulation of humoral theory, after which Asper explains that the word “may, by Metaphore, applie itself / Unto the generall disposition” of a person (lines 95–124). 75See, especially, The Misanthrope, or the Melancholic [Atrabilaire] Lover and The Imaginary Invalid. Hume was surely already familiar with Molièrewhen writing the Treatise. His praise for French theater is expressed in a 1741 essay, published just one year after the Treatise’s completion: “With regard to the stage, they [the French] have excelled even the Greeks, who far excelled the English” (E CL 6, Mil 90–91). The Nortons include Molière’s Oeuvres in six volumes in Hume’s library (Norton and Norton, 1996, 115, No. 876–77). 76On humoral physiology in English literature, see Robin(1911), 19–45; Babb(1951), 1–20; Moore(1953), 181ff. 77On the reception and transformation of mechanistic physiology by the Scottish, see Brown(1968); Guerrini(1985); and Wright(1991b), esp. 255–56.

71 “the Human Body” as a “Machin of infinite Number and Variety of different Channels and Pipes, filled with various and different Liquors and Fluids.”78 By itself, an understanding of Hume’s medical context can simply lead us to think that Hume himself was out of touch with current medicine. But, as interpreters have noted,79 mechanistic views can also be found in Hume’s own writings. Throughout the Treatise’s “anatomy of the mind,”80 Hume’s pervasive talk of the transfer of force and vivacity from impression to idea seems to draw on the notions of particle motion and fluid dynamics so central to mechanistic physiology. Describing this sort of transfer in the case of poetical associations, Hume says that the “vividness. . . is convey’d, as by so many pipes or canals”81—an image reminiscent of Cheyne’s body of “Channels and Pipes.” Moreover, Hume’s brief but repeated appeals to the ‘animal spirits’ in belief, association, and error reflect the brand of Cartesian mechanism which Hume encountered in the works of Malebranche and Mandev- ille. He seems to accept the thesis that the twists and turns of thought correlate with the motions of spirits, subtle fluids coursing through interconnected traces in the brain tissue.82 These passages, and others, demonstrate the great influence of mechanistic physiology on his thought. Perhaps more to the point in a discussion of Treatise 1.4.7, Hume had available non- humoral accounts of the nature and causes of melancholy. He shows his familiarity with one of these in a 1734 letter addressed to an anonymous physician, possibly Cheyne.83 In the letter, Hume complains of a depressive “Distemper” with “repeated Interruptions” of his “Train of Thought,” initially diagnosed as “the Disease of the Learned.”84 As interpreters have noted, this particular label for melancholic or “hypochondrical” disorders likely derives

78Cheyne(1733), 4. 79See Wright(1983), esp. 187–246. 80T 2.1.12.2, SBN 326; cf. 1.4.6.23, SBN 263; 3.3.6.6, SBN 620–21; A 2, SBN 646. 81T 1.3.10.7, SBN 122; cf. DP 6.19, Bea 29. 82See T 1.2.5.20, SBN 60–61; 1.3.8.2, SBN 98–99; 1.3.10.9–10, SBN 123; 1.4.1.10, SBN 185; 1.4.7.10, SBN 269–70; 2.1.1.1, SBN 275; 2.1.5.11, SBN 289–90; 2.2.8.4, SBN 373–74. For discussion, see Wright(1983), 187–246, esp. 190–91, 214–19; Frasca-Spada(2003); Kail(2008), 66–67, 74. 83Burton(1846), 42 lists Cheyne as the intended recipient, as does Greig(1932/2011)), viii, 12n2. Mossner (1944) disagrees, suggesting instead John Arbuthnot. I ﬁnd the rejoinders to Mossner’s arguments in Wright (2003) convincing. Since the discussion below does not depend on the letter’s intended recipient, I leave my own arguments in favor of Cheyne for another time. 84HL 3.14–16.

72 from Mandeville’s 1715 A Treatise of Hypochondriack and Hysterick Passions.85 That text favorably repeats mechanistic views, while mocking Galenism.86 According to Mandeville, melancholy occurs when “the labour of the Brain [has] exhausted. . . the finest Spirits.”87 Hume seems to employ this mechanistic accoubt in his letter. He compares and contrasts his “present Condition” with a “Want of Spirits,” describes philosophical studies as “wasting” his spirits, and asks whether his “Spirits [will ever] regain their former Spring and Vigor.”88 This strongly suggests that Hume had a mechanistic account of melancholy available, and indeed one that was more in line with his scientific context. Why insist that it is Hume who is out of touch, and not the humoral reading? I think neither is out of touch. Despite the rise of newer physiologies, humoral theory continued to influence the medicine of Hume’s time. As medical historians have noted, humoral theory’s decline was “slow, one of evolution rather than revolution,” and still incomplete.89 Amidst all the innovation, Galenism remained surprisingly entrenched in Scottish medical practice. Early eighteenth century Scottish physicians and surgeons continued to diagnose and treat symptoms according to a humoral scheme, prescribing traditional remedies like blood-letting and purges. Their familiarity with humoral theory was reinforced by the medical education available. Before the founding of Edinburgh’s medical school in 1726, aspiring Scottish physicians traveled to mainland Europe, studying at the more affordable universities of Padua, Reims, and, especially, Leiden. Many of the university courses they

85Wright(1983), 216ff, 236n10. Mandeville’s writings were among Hume’s early sources. In the Treatise, Hume portrays Mandeville as a champion of experimental philosophy (T Intro.7n1, SBN xvii; A 2, SBN 646). 86See Mandeville(1715), 38, 108–10; and 11–14, 80, 188–89, respectively. 87Mandeville(1715), 149. 88HL 3.13, 14, 18. 89Porter(1987), 47. As Porter points out, many humoral ideas, and some humoral terms, survived the shift to the newer physiological frameworks, often by being partially reduced to, explained by, or made to serve elements of the latter. As he puts it, many proponents of “the new [physiological] models joined in denouncing Aristotelianism and Galenism as empty and barren—if often in reality recycling their ideas under a different guise, pouring old wine into new bottles” (Porter, 2004, 54). Jackson(1978), 369–73 gives an excellent example from the iatrochemistry of Thomas Willis. Though announcing in his 1672 De anima brutorum that “we cannot yield to what some Physicians affirm, that Melancholy doth arise from a Melancholick humor, somewhere primarily and of itself begotten,” Willis nonetheless speaks of “yellow Bile of Choler. . . in the Gall-Bladder, or the black Bile so called, of Melancholick humor in the Spleen” and preserves many humoral ideas (Willis, 1683, 189, 192–93). These include the idea of a murky fluid stored in the spleen; that overabundance of this fluid causes certain changes throughout the body, leading eventually to irregular movements of ‘spirits’ in the brain; and that these movements underly or simulate the erratic thoughts and passions of a melancholic break.

73 attended consisted primarily in descriptive lectures on classical texts, including broadly humoral treatises from antiquity. At Leiden, for example, MD candidates were required to expound several Hippocratic aphorisms as a part of their examinations.90 As a result, most Scottish physicians practicing during Hume’s youth stocked their libraries with the medical classics, especially Hippocratic texts.91 Even the mechanistic medical texts of Hume’s contemporaries sometimes preserved or revived key elements of humoral theory. Cheyne is a good example. His 1724 Essay of Health and Long Life focuses on the management of the classic six “non-naturals”: “1. The Air we breath in. 2. Our Meat and Drink. 3. Our Sleep and Watching. 4. Our Exercise and Rest. 5. Our Evacuation and their Obstructions. 6. The Passions of our Minds.”92 This six-part division of the regimen structures the entire text. And Cheyne concludes with a call for balance strongly reminiscent of the Hippocratic doctrine of health as proper proportion, urging his readers to “observe the golden Mean in all their Passions, Appetites and Desires.”93 Similarly, Cheyne’s 1733 treatise on melancholy, The English Malady, revived the Hippocratic notion of a national temperament, shaped by atmospheric conditions and cultural mores, and involving a proneness to certain diseases—in England’s case, a variety of “nervous Distempers.”94 Even if Cheyne’s mechanism abandoned the schema of four humors, his medicine still preserved important elements from the humoral tradition. These include the idea of health as a kind of balance; the idea that health is best maintained and restored through the management of the six non-naturals; and the idea that atmospheric and geographical conditions shaped national characters and diseases. Ultimately, the humoral reading, as I deﬁned it, does not require commitment to the underlying physiology of the four humors. It attributes to Hume the view that health requires a balance of the temperaments of humoral theory. This conception of health was by no means

90This practice continued beyond the turn of the eighteenth century, even after Hermann Boerhaave’s inﬂuence brought to medical pedagogy a greater focus on the practical dimensions of diagnosis and treatment. Eventually, Boerhaave’s own aphorisms were used for ﬁnal examinations. But neither of these changes challenged the status of the medical classics in Leiden’s course of study. Indeed, Boerhaave’s inaugural lecture was a panegyric for the study of Hippocrates. See Dingwall(1995), 101–02, 169; Porter(1999), 290–91. 91Dingwall(1995), 145. 92Cheyne(1724), 2. 93Cheyne(1724), 120. 94Cheyne(1733), i–ii. On national temperament in Hippocrates, see note 21 above.

74 long gone in Hume’s day, and it is not antiquarian to suggest that he took it seriously. Though certainly in decline, humoral medicine still had some currency in Hume’s Scotland. In this respect, the anachronism objection is itself anachronistic. Still, the objection raises an important question that has so far not been answered. Did Hume in fact believe in the underlying physiology of the four bodily humors? The evidence of his mechanism can leave a lingering doubt about the extent to which he accepted humoral theory. To answer this question, I turn next to Hume’s views about the role of physiology in his philosophical method.

3.4 THE METHODOLOGY OBJECTION

The humoral reading can seem fundamentally at odds with Hume’s experimental method in the Treatise. For Hume seems to express a methodological commitment to avoid appealing to any speciﬁc theory of human physiology. When leaving aside certain passions in the Treatise’s second book, Hume says: “the examination of them wou’d lead me too far from my present subject, into the sciences of anatomy and natural philosophy.”95 The same commitment seems to underlie Hume’s introduction of “impressions of sensation” as “aris[ing] in the soul originally, from unknown causes.”96 By choosing not to specify the causes of such impressions, Hume can seem to relegate the laws by which they enter the mind to natural philosophy, excluding them from philosophy proper. His doing so can seem to suggest that he thinks such laws would fail to give properly philosophical explanations of mental phenomena.97 And that would seem to suggest that Hume would bar physiological theories from doing explanatory work in his philosophy. If that is right, and Hume stays faithful to this methodological commitment, then he cannot appeal to humoral theory to explain the succession of impressions and ideas in 1.4.7. I call this the methodology objection. One possible reply is to deny that Hume is fully faithful to this methodological commit-

95T 2.1.1.2, SBN 275–76. These are the immediate passions which “without introduction make their appearance in the soul,” and so “depend upon natural and physical causes.” 96T 1.1.2.1, SBN 7, my italics. 97See Broughton(2008), 43–44; Alanen(2008), 184–85.

75 ment, if he holds it at all. For Hume seems to give numerous, straightforwardly physiological explanations of mental phenomena. He says, for example, that an “extraordinary ferment of the blood and spirits” is a cause of madness, and a “new tide” of blood interrupts melancholy.98 In at least one passage, Hume explicitly announces that he “must. . . have recourse” to physiology to explain the mental phenomenon at issue. There, he is trying to “account for the mistakes that arise from [the] relations” of resemblance, contiguity, and causation. In other words, Hume wants to explain why, when thinking about some object, we replace it, without being aware that we are doing so, with another object that only resembles the ﬁrst, or that we previously experienced nearby or in close succession with the ﬁrst. To do so, he conducts “an imaginary dissection of the brain”:

I shall. . . observe, that as the mind is endow’d with a power of exciting any idea it pleases; whenever it dispatches the spirits into that region of the brain, in which the idea is plac’d; these spirits always excite the idea, when they run precisely into the proper traces, and rummage that cell, which belongs to the idea. But as their motion is seldom direct, and naturally turns a little to the one side or the other; for this reason the animal spirits, falling into the contiguous traces, present other related ideas in lieu of that which the mind desir’d at ﬁrst to survey. This change we are not always sensible of; but continuing still the same train of thought, make use of the related idea, which is presented to us, and employ it in our reasoning, as if it were the same with what we demanded. This is the cause of many mistakes and sophisms in philosophy; as will naturally be imagin’d, and as it wou’d be easy to shew, if there was occasion.99

On the psychophysiological theory of mind Hume invokes, each idea is “plac’d” in a region of the brain. When the mind is about to “excite” a certain idea, it “dispatches” the spirits, sending them along traces in the brain tissue toward a “cell” which houses that idea. If the spirits reach the cell, we form the idea. But since the flow of the spirits is naturally a bit erratic, the spirits divert into nearby traces, and so arrive at different cells and lead us to form different ideas. So far, this falls short of a complete explanation of why the ideas we end up with bear relations of resemblance, contiguity, or causation to the first idea—the one toward whose cell the spirits were first “dispatched.” For that, we would need a further story about why ideas which bear these relations are located on contiguous brain traces. Hume does not give this story. He instead seems to presuppose his readers’ familiarity with

98T 1.3.10.9, SBN 123; 2.2.4.4, SBN 352–53. 99T 1.2.5.20, SBN 60–61.

76 it—that his talk of ‘spirits,’ ‘cells,’ and ‘traces’ would be sufficient to prompt his readers to fill in the relevant details.100 We need not concern ourselves with the details here. The passage is of interest insofar as it seems to be a stark example of Hume reaching for a physiological explanation. It is true that Hume apologizes for the explanation he gives here, noting that it is in conflict with the “first maxim” of his philosophy. But that maxim does not say or imply that he must avoid physiological explanation. It instead says “that we must rest contented with experience” as the test of our judgments. Hume’s “dissection” violates that maxim because the physiology he appeals to is “an imaginary dissection of the brain,” rather than one drawn from experience.101 In short, he apologizes not for his explanation’s being physiological, but for its being speculative. This passage, and others, suggest that Hume does invoke physiological explanations after all. Their prevalence may lead us to question whether it is really a feature of Hume’s methodology to avoid appeals to physiology. Doubting that Hume holds such a commitment could perhaps motivate a reading of the seemingly methodological passages as merely delineating the topic of Hume’s inquiries. On that reading, Hume does not restrict what sorts of things can figure in his explanations, but rather what sorts of things he is to explain in the first place. He relieves himself of the obligation to give explanations that would take him deep into the details of human physiology. And he is prudent to do so, since such explanations would require “experiments” of a very different sort than his “cautious observation of human life. . . in the common course of the world, by men’s behaviour in company, in affairs, and in their pleasures.”102 But this need not bar physiological explanations from appearing in Hume’s theory of the mind altogether. Perhaps some topics, like that of the sentiments and ideas which appear in a mind recovering from melancholy, may benefit from drawing on a rich tradition of physiological explanation. I think this reply to the methodology objection holds water. But Hume’s “imaginary dissection” also provides a second and at least as compelling reply. For his discussion of the “dissection” suggests a general lesson about why Hume bothers to state the physiological

100He would not be presumptuous to expect this. Such a story was implicit in a common seventeenth and early-eighteenth century view of the ‘imagination,’ which conceived it as a physiological structure in which sense impressions formed traces. Wright(1983), 188–92 gives a good overview. 101T 1.2.5.20, SBN 60. 102T Intro.10, SBN xix.

77 causes of some of his topics of interest. Just before conducting the “dissection,” Hume says something very curious:

I shall only premise, that we must distinguish exactly betwixt the phænomenon itself [making mistakes across the three relations], and the causes, which I shall assign for it [the diversion of animal spirits across contiguous brain traces]; and must not imagine from any uncertainty in the latter, that the former is also uncertain. The phænomenon may be real, tho’ my explication be chimerical. The falshood of the one is no consequence of that of the other.103

Here, Hume tells us that his own physiological explanation for the relevant mistakes in reasoning might be “chimerical.” His caution here would be in keeping with his Scottish Newtonian predecessors, who took animal spirits to be an antiquated fiction.104 Nonetheless, Hume reassures his reader that the “falshood” of the causal explanation he is about to offer would not imply the falsehood of the “principle” he is trying to explain. But this raises the question: What could be the point of invoking an admittedly dubious physiological cause? What does Hume hope to gain by doing that? The answer, I think, is that citing a physiological cause invokes a larger physiological story or framework, which serves as a rich source of analogy for thinking about the mind. Such an answer is suggested by remarks earlier in the Treatise. Here, Hume discusses the importance of analogy to the “explication” of a certain mental phenomenon at issue—in this case, how our particular ideas can become general in their representation. He says: “To explain the ultimate causes of our mental actions is impossible. ’Tis sufficient, if we can give any satisfactory account of them from experience and analogy.”105 Hume’s imaginary dissection does not provide us with an account from experience, since it is imaginary. But Hume’s story about diverted spirits, and the story about the formation of brain traces which Hume expects his readers to fill in, together with the general framework of fluid dynamics which both stories share, provide rich analogies for understanding the kind of error at issue. In this case, spatial and mechanical notions offer a way to think about how the mind makes its unnoticed substitutions of ideas across the three relations. Whether or not the stories 103T 1.2.5.20, SBN 60. 104Cheyne(1733), 89, for instance, criticized the animal spirits for being “of the same Leaven with the substantial Forms of Aristotle and the celestial System of Ptolemy.” 105T 1.1.7.11, SBN 22.

78 have really got our physiology right is not crucial for Hume’s enterprise. One or more of the stories might turn out to be false, or impossible to judge, but no matter. The analogies they offer would still, in that case, give anyone familiar with the stories and frameworks a way of thinking about the mind—a way which could illustrate the “principle” at issue, and bring about a kind of satisfaction. If that is right, Hume need not commit himself to the correctness of the exact physiological stories he gives. In the case of mechanistic explanation, his imaginary dissection does not require a belief in the existence of animal spirits. This may reveal a limited sense in which the methodological objection is on point: From time to time, Hume evinces wariness about committing himself to the literal truth of this or that physiological theory, even while he is willing to state its theses. As a result, it is often unclear from the text whether Hume endorses the physiological theses he states, even half-heartedly. But his level of endorsement does not always matter for his purposes. Even someone who rejects those theses could make room for the analogical use of physiological theses and frameworks, especially if she expected her audience to have prior familiarity with them. We can now see that the humoral reading need not attribute endorsement of humoral physiology to Hume. In the end, I think it is genuinely unclear whether, or to what extent, Hume believed in humoral physiology. But it does not matter. For the humoral reading can interpret Hume’s talk of bodily humors as meant to offer analogies for understanding the workings of the mind. Presumably, such analogies would suggest that the mind, like humoral theory’s picture of the body, has four basic elements which, when balanced, constitute health or proper functioning. But that is precisely the core of humoral theory’s psychological part. The primary resource Hume borrows from humoral theory, then, is its robust temperament psychology. When first expounding the humoral reading in §3.2, I described the progression of stages in 1.4.7 in largely physiological terms: an excess of black bile, dissolved by a “new tide” of blood, expunged by yellow bile, cooled by phlegm. Doing so helped to make perspicuous the various oppositions between the stages, and thus the logic behind their succession. But we can now take this language to be analogical, and so read Hume’s progression as modeled only on humoral psychology: melancholy, invigorated by sanguinity, scorned by biliousness,

79 calmed by phlegm. We can then summarize the humoral reading ﬁnally in terms of a succession of temperaments:

Hume’s stages in 1.4.7, with corresponding temperaments melancholy → sociability → aggression → composure → curiosity melancholic → sanguine → bilious → phlegmatic → even-tempered

This would be a humoral reading which jettisons the underlying physiology of gross bodily ﬂuids—a reading on which healthy mental functioning consists in the balance of four basic psychological temperaments, bearing the same relations of opposition as their physiological counterparts, each having tendencies to cause or be caused by characteristic sentiments and activities. That would still be a ripe source of explanation. In summary, the methodology objection objects that Hume would bar physiological theories, such as that of humoral medicine, from doing explanatory work in his philosophy. The ﬁrst reply I considered was that Hume includes a range of physiological explanations throughout the Treatise. The second reply was that Hume’s apparent invocations of humoral physiology can be read as analogies, rather than as literal endorsements, drawing on the precedent set by Hume’s remarks on the use of analogies in explaining the mind. One reply denies Hume’s exclusion of physiology; the other accommodates it. In either case, Hume’s invocation of the humors is consistent with his actual method. Either reply is enough to answer the methodology objection. And I am inclined to think both are correct. Hume makes apparently physiological claims throughout the Treatise, and 1.4.7 is no exception. And these claims are a rich source of analogy, whether or not Hume literally believes them. Either way, the stages of 1.4.7 can be seen as a cycle through the four temperaments en route to psychic balance.

3.5 RECONCEIVING HUME’S SKEPTICISM

The four humors or temperaments, together with their characteristic activities and sentiments, explain Hume’s succession of stages in 1.4.7. They also provide insight into his views on skepticism. Hume’s association of skepticism with the melancholic humor or temperament

80 encourages us to rethink both what he thinks skepticism is, and what role he takes it to play in human life. First, the association suggests that skepticism itself can be understood as a temperament, or disposition to think, act, and feel in certain ways, which can be more or less dominant, and expressed to greater and lesser degrees. The melancholic temperament, which tends to inspire engagement in inquiry, caution in forming beliefs, and a felt doubt or unease concerning one’s limitations, is a natural candidate for the tendency of mind Hume labels “scepticism.” When such a temperament dominates, unchecked by the other temperaments, one loses one’s grip on reality and sinks into a gloomy “delirium.” But when the temperament is balanced with others, it can contribute the carefulness, focus, and awareness of one’s abilities needed for sober and scientiﬁc pursuits.106 The humoral reading alone does not force us to see Hume as viewing skepticism as a temperament. It only suggests this interpretation. But the suggestion is supported by the language Hume uses when discussing skepticism. In the Treatise, he speaks of a “sceptical disposition,”107 and often uses the word ‘disposition’ in close proximity and interchangeably with ‘temper.’108 Hume acknowledges that this disposition or temper admits of varying degrees of intensity. In the Appendix to the Treatise, for example, he speaks of “a modest scepticism to a certain degree.”109 And in Treatise 1.4.3, he speaks of the “true philosopher”

106In the essay “Of Superstition and Enthusiasm,” Hume claims that superstition arises “from a gloomy and melancholic disposition,” among other causes (E SE 2, Mil 73). This can seem to threaten the idea that Hume associates the melancholic temperament with carefulness, since superstitious people are not careful in forming beliefs. “In such a state of mind,” Hume says, “infinite unknown evils are dreaded from unknown agents; and where real objects of terror are wanting, the soul, active to its own prejudice, and fostering its predominant inclination, finds imaginary ones, to whose power and malevolence it sets no limits” (E SE 2, Mil 74). But the melancholic temperament’s being a cause of uncautious superstition does not imply that it is not characterized by caution. Compare The Natural History of Religion, where Hume says that “the mind, sunk into diffidence, terror, and melancholy, has recourse to every method of appeasing those secret intelligent powers, on whom our fortune is supposed entirely to depend” (NHR 3.4, Bea 42). Here, even “diffidence”—the opposite of confidence—is a cause of superstitious belief in invisible and unknown beings. If anything, this passage suggests a close relationship between melancholy and diffidence, carefulness, or doubt. There is still a further question about why these qualities should lead into superstition. I speculate that Hume’s answer would be that such qualities can, in their excess, make one feel weak, and accordingly vulnerable, even to invisible powers. But whatever Hume’s answer may be, the tendency of melancholy, and its associated carefulness or doubt, to mutate into superstitious madness would have already been familiar to him through the humoral tradition. 107T 1.4.7.10, SBN 269. 108See T 2.2.4.6, SBN 354; 2.3.8.13, SBN 437; 3.3.2.3, SBN 593; cf. 3.2.2.12, SBN 481. 109T 1.2.5.26n12, SBN 639.

81 as embodying a “moderate scepticism.”110 These modest and moderate degrees contrast with an extreme skepticism, which Hume labels “total”111 and “extravagant.”112 We can read both of the latter labels as modiﬁers that express the intensity or extent of the skeptical temperament: ‘Total,’ on this reading, connotes reaching a maximum, total dominance; ‘extravagant’ connotes going too far, a lack of moderation, indulgence. Now, such a reading allows that “total skepticism” can at times refer to the scope of a doubt—to skepticism about all of our beliefs. But this is not the only way in which Hume conceives of skepticism as being or becoming “total.” In the section “Of scepticism with regard to reason” (1.4.1), he writes of an extreme loss of conﬁdence:

When I reﬂect on the natural fallibility of my judgment, I have less conﬁdence in my opinions, than when I only consider the objects concerning which I reason; and when I proceed still farther, to turn the scrutiny against every successive estimation I make of my faculties, all the rules of logic require a continual diminution, and at last a total extinction of belief and evidence.113

The melancholic temperament involves a disposition to reason, often repetitively or obses- sively, and often with regard to one’s own weakness or fallibility. One who, like Hume in this passage, does not just reflect on her fallibility once, but “proceeds still farther, to turn this scrutiny against every successive estimation” could be seen, at that moment, as exhibiting an extremely skeptical temperament—indeed, a temperament that has become so ‘total’ that it now entirely characterizes her mental landscape, eclipsing all opposing dispositions to judge and feel, and so destroying her confidence in her beliefs. We may read the ‘total scepticism’ of 1.4.1, then, as the complete domination by a skeptical temperament. As Cleanthes puts it in Hume’s Dialogues Concerning Natural Religion, “total scepticism” arises “in a flush of humour.”114 The language of varying degrees is a point of continuity with Hume’s discussions of skepticism in the first Enquiry.115 In Section XII, he discusses four varieties of skepticism,

110T 1.4.3.10, SBN 224. 111T 1.4.1.7, SBN 183; 1.4.7.7, SBN 268. 112T 1.4.2.50, SBN 214; 1.4.4.6, SBN 227–28; cf. EHU 12.17, SBN 155. 113T 1.4.1.6, SBN 183. 114DNR 1.6, KS 132–33. 115There are still many important diﬀerences between the discussions of skepticism in the Treatise and the ﬁrst Enquiry. For a comprehensive discussion, see Qu(2020).

82 grouped into two pairs. Strikingly, within each pair one variety is excessive, and a threat to reasoning and inquiry, while the other is moderate and useful. Hume’s main concern in this section is the two varieties of skepticism which are “conse- quent to science and enquiry.” These are adopted only after discovering the “absolute falla- ciousness” or “unﬁtness” of our mental faculties for their common and speculative uses.116 One of these varieties Hume calls “excessive scepticism,” and associates with the Pyrrhonian skeptics of ancient times.117 The label ‘excessive’ suggests that the skepticism has been taken too far or become too dominant. It is characterized by an attempt to preserve a widespread doubt and sense of “universal perplexity and confusion.”118 The other variety Hume calls “mitigated scepticism,”119 and associates with the academic skeptics. ‘Mitigated’ suggests a lessening of intensity through interaction with other forces—hence, it suggests a kind of balance. And, indeed, Hume says that mitigated skepticism just is the excessive skepticism “in some measure, corrected” by other tendencies of the mind. The result is a kind of “modesty and reserve.”120 The other two varieties of skepticism Hume discusses are “antecedent to all study and philosophy” and are meant to serve as “preservative[s] against error and precipitate judgment.”121 The ﬁrst of these, which Hume associates with “Des Cartes,”

recommends an universal doubt, not only of all our former opinions and principles, but also of our very faculties; of whose veracity, say they, we must assure ourselves, by a chain of reasoning, deduced from some original principle, which cannot possibly be fallacious or deceitful.122

According to Hume, such a skepticism would be a poor preparation for philosophy, for it would leave us in perpetual suspension of judgment. By continually declining to assent to any judgement until we have assured ourselves of the veracity of our faculties, we eﬀectively

116EHU 12.5, SBN 150. 117EHU 12.23, SBN 159; 12.24, SBN 161; cf. 12.21, SBN 158–59. 118EHU 12.24, SBN 162 119EHU 12.24–25, SBN 161. 120EHU 12.24, SBN 161. Fogelin(2009), 6, cf. 158, also takes a balance between tendencies or mechanisms of the mind to be central to mitigated skepticism: “When the destructive mechanisms of Pyrrhonism—Hume’s label for radical skepticism—are counterbalanced by the mechanisms that produce common (vulgar) belief, then the mind, as a result of the vector of these two opposing forces, naturally settles into the standpoint of a mitigated or moderate skepticism.” 121EHU 12.3, SBN 149. 122EHU 12.3, SBN 149–50.

83 deprive ourselves of assenting to any “principle” which could provide that assurance. As a result, “no reasoning could ever bring us to a state of assurance and conviction upon any subject.”123 In the next paragraph, however, Hume acknowledges that the very same “species” of skepticism,

when more moderate, may be understood in a very reasonable sense, and is a necessary preparative to the study of philosophy, by preserving a proper impartiality in our judgments, and weaning our mind from all those prejudices, which we may have imbibed from education or rash opinion.124

This “more moderate” variety still advises us “[t]o begin with clear and self-evident principles.” But it does not demand that we show, through a chain of reasoning, that such principles can never deceive. For a skeptic of this moderate sort, it is enough to embody a degree of caution—“to advance by timorous and sure steps, to review frequently our conclusions, and examine accurately all their consequences.”125 In adopting antecedent skepticism, then, one develops a propensity to step back and withhold any assent which is too quick and easy. This may involve checking the steps of one’s reasoning, examining before judging, making sure one has not made a mistake or imported a bias, and, when possible, looking for firmer foundations. But the “Cartesian” variety takes this to excess. One who goes so far can never find a foundation firm enough and so ends up in a continual search. None of this is to deny that Hume sometimes uses the term ‘scepticism’ to refer to a particular, historical, philosophical sect or tradition or its core principles. Hume clearly uses the term this way at points. In the first Enquiry, for example, Hume seems to regard skepticism as a body of “philosophical principles”126 that can be supported by argument or reasoning: He says that skeptics give “profound arguments against the senses”127 and “philosophical objection[s] to the evidence of sense,”128 and that they “attempt. . . to destroy reason by argument and ratiocination.”129 In the Treatise, he speaks of “sceptics” as a “fantastic sect”

123EHU 12.3, SBN 150. 124EHU 12.4, SBN 150, my emphasis. 125EHU 12.4, SBN 150. 126EHU 12.2, SBN 149; see also 12.21, SBN 158–59; 12.23, SBN 160. 127EHU 12.6, SBN 151. 128EHU 12.16, SBN 155. 129EHU 12.17, SBN 155.

84 that offers up “cavils” for their “opinions” about the “uncertainty” of our judgments.130 At first, a conception of skepticism as a temperament seems ill fit to accommodate these uses. A temperament is not a “philosophical principle,” or body thereof, or a tradition or method. And though a temperament could perhaps be adopted or inculcated as the result of reasoning, it cannot itself be the conclusion of an argument. But a temperament can be the source of the reasoning or tendency which leads one to adopt a body of principles. Hume tells us as much when he claims that the “decisions” and “pursuits” of “almost every” philosopher are dictated by his “predominant inclination, to which his other desires and affections submit, and which governs him, though, perhaps, with some intervals, through the whole course of his life.”131 In this way, a temperament can stand behind, shape, and lend character or content to a body of principles. The “sceptical. . . system of philosophy” referred to in the title of the fourth part of the Treatise’s first book can then be read as a collection of arguments or principles that a person of a strongly inquisitive temperament would be apt to make. Such a collection could include within it a long philosophical tradition insofar as the tradition resulted from the same predominating temperament—one of continual questioning, repeated applications of reason, and the doubt which follows on their coattails. And Hume can call his philosophical exploration of the human understanding “sceptical” insofar as its repeated application of causal reasoning “tends to give us a notion of the imperfections and narrow limits of human understanding.”132 The idea that a philosophical sect or system can be characterized by a predominating temperament can seem foreign. But in viewing the skeptical system in this light, Hume in effect takes a more classical perspective on the character of philosophical traditions. That perspective views them primarily in terms of competing ways of living, and secondarily in terms of the arguments or principles which could be used to justify those lived dispositions. In that light, the ‘skeptikos’ is first and foremost just what that word means: a person who

130T 1.4.1.7–8, SBN 183. 131E Sc 1–2, Mil 159–60. See also EHU 5.1, SBN 40. Compare Nietzsche’s Gay Science, which Nietzsche says is “written in the language of the wind that thaws ice and snow: high spirits, unrest, contradiction, and April weather.” This “science” is an unruly collection of observations and prescriptions which belong together insofar as they arise from a prevailing, albeit temporary, mood—“the intoxication of convalescence” (Nietzsche, 1974, 32). They form an “art. . . of cheerfulness,” which betrays “a playful tenderness” (37). 132A 27, SBN 657.

85 ‘scopes out,’ observes, considers, examines, or inquires.133 Though the historical sect may have, as a matter of fact, been characterized by an excess in these activities, a more moderate kind is possible. As we have seen, Hume characterizes skepticism both as a temperament and as a doctrine, rather than only as a temperament. But the temperamental conception appears to be primary. Hume uses the skeptical temperament to explain the appearance of skeptical doctrines. And a preponderance of temperamental language reveals a widespread emphasis on the temperamental conception. This conception finds textual support in Hume’s repeated invocation of the humors, his talk of skepticism as a disposition, his emphasis on varying degrees of skepticism and on its moderation or excess, and his association of skepticism with caution. These are not easily accommodated by a conception of skepticism as a doctrine or theory. The conception of skepticism as a temperament also allows us to resolve a puzzle about skepticism’s role in Hume’s theory of human nature. For Hume seems to say conflicting things. He calls skepticism a “malady,”134 and describes its unpleasant symptoms in great detail. But he also treats it as an aid, prescribing “tinctures” of it to “abate” certain ailments,135 and saying: “[i]n all incidents of life we ought still to preserve our scepticism.”136 How can a single phenomenon, skepticism, be both disease and medicine? The likening of skepticism to the melancholic temperament of humoral theory offers an answer. A humor or temperament has varying degrees of intensity; it can both dominate, destroying healthy psychic balance, and be moderate, restoring or maintaining this balance. Hume can call skepticism a “malady,” because it is harmful when it is overabundant and grips the mind with too much intensity. And Hume can treat skepticism as a medicine: In appropriate doses,

133Compare Livingston(1998), 7–11, which discusses the eudaemonistic sects and Hume’s self-identification with skepticism. My reading of Hume’s skepticism perhaps invites a comparison with Hellenistic skepticism, insofar as my reading privileges a notion of balance, which helps to combat dogmatism and make one carefree. At first glance, this is reminiscent of Pyrrhonism. Pyrrhonians thought that a certain philosophical method, which brought about a “balance” (isosthenia) between diametrically opposed appearances, theories, or arguments, could counteract the tumults of dogmatism, and bring about tranquility and quietude of the soul. But the comparison quickly breaks down. The balance I find in Hume’s discussions of skepticism is not a feature of a philosophical method—even if a recovery from excessive philosophy can help inculcate it. It is not a balance between opposing theories, but between opposing temperaments. 134T 1.4.2.57, SBN 218. 135EHU 12.24, SBN 161. 136T 1.4.7.11, SBN 270.

86 usually small, it can temper excesses in the other humors or temperaments, and contribute its share of the dispositions that together constitute a healthy mind. Hume describes this function in the ﬁrst Enquiry:

If any of the learned be inclined, from their natural temper, to haughtiness or obstinacy, a small tincture of Pyrrhonism might abate their pride, by shewing them, that the few advantages, which they have attained over their fellow, are but inconsiderable, if compared with the universal perplexity and confusion, which is inherent in human nature.137

Hume does not specify which humor or temperament characterizes the “natural temper” he has in mind.138 But he comes closer in Treatise 1.4.7, where he describes the kind of person who can benefit from skepticism as having a “warm imagination.” Someone with a warm imagination, Hume says, dogmatically embraces philosophical “hypotheses. . . merely for being specious and agreeable.”139 If uneducated, such a person is prone to a kind of religious frenzy or raving which Hume calls ‘enthusiasm.’ For Hume, warmth of the imagination proceeds from “luxuriant health” and “a bold and confident disposition”140—all of which are characteristic of the temperament associated with excessive blood. We can thus conclude that Hume attributes obstinacy and dogmatism to excessive sanguinity. An overly sanguine temperament, he thinks, can be treated by a brief experience of skeptical philosophy. A short glimpse of the depressing state of our cognitive faculties can contribute the “degree of doubt, and caution, and modesty, which, in all kinds of scrutiny and decision, ought for ever to accompany a just reasoner.”141 Hume can then be read as offering the skeptical strands of his own philosophy as the preferred melancholic tonic. Because such a tonic can treat other humoral or temperamental excesses—especially the unbridled enthusiasm and dogmatism of excessive blood or sanguinity—it belongs within our arsenal of cures as human beings interested in living bal-

137EHU 12.24, SBN 161. 138In fact, Hume does not explicitly name any humors in his discussion of skepticism at the end of the ﬁrst Enquiry. Nonetheless, as we have seen, he still treats skepticism as a temperament which may have various degrees of intensity. Moreover, humoral words and allusions do appear in his prior discussion of the role of philosophy in human life. There, he continues to associate “melancholy” with excessive philosophy. And he continues to view a healthy life—one which “nature has pointed out. . . as most suitable to the human race”— as including some philosophy, so long as it is “mixed” with activities associated with the other temperaments, like socializing and rest (1.6, SBN 8–9). 139T 1.4.7.14, SBN 272. 140E SE 3, Mil 74. 141EHU 12.24, SBN 162.

87 anced, healthy lives. This, I imagine, is Hume’s primary intention in considering skepticism in the context of a recovery from melancholy described in humoral terms. He shows us not only how to emerge from our own intellectual depression, but also how the experience of that depression can be an education—how it can prepare us to cure ourselves and others of other disorders or imbalances. If the humoral reading of 1.4.7 is right, then the common interpretation that Hume rejects a kind of skeptical theory in that section is distorting in two important ways. According to that interpretation, what Hume rejects is a theory of human knowledge on which we can know very little—presumably, a theory implied or inspired by some of Hume’s ﬁndings earlier in the Treatise. In contemporary philosophy, the word ‘skepticism’ often refers to just that: a negative or austere theory of human knowledge adopted on philosophical grounds. It is “the view that we know nothing, or that nothing is certain, or that everything is open to doubt,” or “that we know very little or nothing of what we think we know.”142 But importing this sense of the word into Hume’s thought can obscure the fact that he generally conceives of skepticism at least largely as a temperament. After all, a theory is not a disease, or cure, or disposition. Secondly, Hume does not exactly reject skepticism in 1.4.7, any more than he simply accepts it. Rather, he tempers it. On his view, eradicating skepticism would be unhealthy. Instead, he moderates skepticism, considered as a temperament, by counterbalancing it with our other tendencies of thought and feeling.

3.6 CONCLUSION

At this point, we might wonder: What are Hume’s views about the substantive doctrines associated with skepticism? What should we make of his arguments for or against those

142Stroud(1984), vii; Nozick(1981), 197, respectively. See also Comesa˜naand Klein(2019), §1: “Philo- sophically interesting forms of skepticism claim that we do not know propositions which we ordinarily think we do know.” Various recent interpreters of Hume attribute this theory conception of skepticism to Hume, or else claim that Hume takes pains to reject it. Some of the clearer examples come from the ﬁrst camp. For example, Fogelin(1985), 6: “Hume accepts a theoretical epistemological skepticism.” According to Meeker (1998), 34, cf. 32: “Hume’s system is sceptical in the sense that it denies that humans have knowledge.”

88 doctrines? The humoral reading suggests a way to approach these questions. We can use humoral theory’s conception of health to help determine whether skeptical doctrines arise from a healthy state of mind. We can then ask whether Hume draws a close connection between health and balance on the one hand, and truth and justification on the other.143 If he does, considerations of health and balance would play a crucial role in Hume’s views about what we should believe. I have not attempted to address these further questions here. Instead, I have offered a defense of the humoral reading, and have used it to reconceive Hume’s skepticism and resolve two important puzzles. First, the humoral reading answers our interpretive puzzle about the succession of the stages in 1.4.7. As I have argued, Hume’s pervasive humoral vocabulary and imagery reveal that he views his recovery from melancholy as a cycle through the four humors or temperaments, leading finally to their balance. Second, the humoral reading helps to explain how Hume can see skepticism as both a disease and a cure. It does so by encouraging us to see Hume as viewing skepticism as a temperament, subject to both excess and moderation. Why is this reading novel? Why have Hume’s humoral allusions gone mostly unnoticed? The answer, I suspect, is this: Because humoral theory has long since fallen from favor, contemporary readers are apt to view the allusions as mere artful flourishes, rather than the terms of art they are. Michael Williams calls 1.4.7 “one of the most dramatic expositions of skeptical doubt ever set down.”144 Ainslie goes a step further, imagining it “must be the most literary stretch of writing in the English-language philosophical canon.”145 While this praise may be deserved, I think it has sometimes obscured the extent to which Hume’s choice of words and images in 1.4.7 serves not only his aesthetic ends, but also his philosophical

143These are central questions for debates about Hume’s naturalism. Both are already explored in Kemp Smith(1941), who identifies certain mental tendencies as, on the one hand, “natural to the mind,” “necessary for its proper functioning,” and conducive to “health and equilibrium” (493–99) and, on the other, “adaptive” (76) or “reliable” (382). More recently, interpreters have suggested an answer to the second question, by attributing to Hume a “proper functionalist” epistemology. See Schmitt(1992), 69; Wolterstorff (1996), 166n6; Greenberg(2008), 728–29. Compare Schmitt(2014), 360–75 and 362n32. For concerns see Meeker(2006). The humoral reading can help develop such interpretations, by offering a rich conception of our ‘healthy’ or ‘natural’ condition—namely, as involving a balance of the four classical temperaments. I leave a more detailed discussion of these interpretations for another time. 144Williams(2004), 267. 145Ainslie(2015), 218. Ainslie does draw attention to melancholy. But his literary praise is a good example of what I have in mind, and may make it easier for him to overlook the other humors. See note 16.

89 ones. What can look like lyricism or eighteenth century “charm” is in fact a substantive doctrine. That doctrine is a conception of health as a balance of temperaments, each of which contributes its share of dispositions to reason and believe in certain ways. The melancholic, or skeptical, temperament contributes doubt, caution, and modesty to our reasoning and inquiry. The sanguine temperament contributes hope and confidence. The bilious and phlegmatic, though they receive less discussion, still plausibly contribute resolve and equanimity, respectively.146 In excess, any of these dispositions becomes pernicious: The melancholic grows into widespread doubt and madness; the sanguine into hasty dogmatism and zealotry; the bilious into short-sightedness; and the phlegmatic into dullness. But, when cooperating through mutual checks and balances, they produce a “just reasoner.”147 Sanguine confidence gets inquiry off the ground, while melancholic caution keeps it from soaring too high. Even the “indolence” of the phlegmatic temperament, Hume thinks, contributes a safeguard against tarrying too long in abstruse reflection.148 The core of this conception is separable not only from humoral physiology, but also from the humoral psychology of the four temperaments. Its basic idea is that proper mental functioning involves a balance of tendencies to reason and believe in certain ways, and that certain epistemic vices, such as skepticism and dogmatism, are the extreme expression of the very same tendencies. These vices are then more a matter of degree than of doctrine. As a result, even skeptics and dogmatists can lead us toward proper mental functioning, when adopting some share of their dispositions helps us correct our own imbalances. We do not need to accept humoral theory, or even its psychological portion, in order to appreciate or accept this idea. Nonetheless, it is an idea that Hume’s invocations of humoral theory can lead us to see for the first time. So is Hume, then, offering a kind of ‘virtue epistemology’? Hume’s use of humoral theory suggests a way in which it is not misleading to say that Hume thinks of skepticism in terms of virtue: a skeptical temperament, to the right degree, contributes to excellence in reasoning

146It is, after all, when Hume is bilious that he “resolve[s] never more to renounce the pleasures of life for the sake of reasoning and philosophy” (1.4.7.10, SBN 269, my emphasis). 147EHU 12.24, SBN 162. 148EHU 1.12, SBN 12.

90 and believing. But this is not the broadly pragmatic conception of virtue we ﬁnd later in the Treatise, centered on the usefulness and agreeableness of a character trait. It is, instead, the more typically Aristotelian picture of proper functioning, and especially of balance, that plays the decisive role in Hume’s conception and treatment of skepticism.149 This balance is a mean with respect to various dispositions, each of which tempers the others. It is in this context that skepticism emerges for Hume as a stage, a temperament, a malady, and a cure.

3.7 TRANSITION: JUST A PALLIATIVE?

As I have argued, Hume viewed his own skeptical philosophy as a potent medicine. It could be used to challenge and moderate an excessive confidence he found common to dogmatic philosophers and religious enthusiasts. Insofar as Hume thought “generally speaking, the errors in religion are dangerous; those in philosophy only ridiculous,” he seemed to have viewed skeptical philosophy as having a greater utility in curbing the religious enthusiast’s extravagances.150 Still, in both cases, Hume hoped that his philosophy could help instill “a correct Judgement,” which “avoiding all distant and high enquiries, confines itself to common life, and to such subjects that fall under daily practice and experience, leaving the more sublime topics to the embellishments of poets and orators, or to the arts of priests and politi- cians.”151 “To bring us to so salutary a determination, nothing can be more serviceable,” he thought, than “to be once thoroughly convinced of the force of the Pyrrhonian doubt, and the impossibility, that any thing, but the strong power of natural instinct could free us from it.”152 He thought that, after such an experience, philosophers, once their inclination to their profession returns through natural causes, would “reflect, that. . . philosophical decisions are nothing but the reflections of common life, methodized and corrected” and would “never [again] be tempted to go beyond common life, so long as they consider the imperfections of

149Contra Schafer(2014), who reads 1.4.7 as “articulating a distinctively Humean account of epistemic virtue” which, “like Hume’s account of moral virtue” in the Treatise’s third book, “is rooted in our passionate nature” (4)—that is, in the passions we naturally feel in response to the utility and agreeableness of character traits. 150T 1.4.7.13, SBN 272. 151EHU 12.25, SBN 162. 152EHU 12.25, SBN 162.

91 those faculties which they employ, their narrow reach, and their inaccurate operations.”153 Apart from naturally tempering sanguinity, I think Hume leaves it somewhat unclear why holding the skeptical strands of his or others’ philosophy in mind has the effect of constraining our thought to the sphere of common life. Perhaps adopting Hume’s view that the relation of cause and effect is the primary function through which the mind forms beliefs about objects not immediately present to the senses, and that this function depends on a habit arising from experience of past conjunctions of events presumably encountered within the common affairs of life, could go some way toward explaining this constraint.154 But these are different grounds for constraint than the experience of skeptical doubt followed by a natural recovery. And it is unclear how even the former grounds could prevent, for example, superstitious belief in imagined beings, insofar as those beings seem, to an overly superstitious mind, to present themselves vividly in experience. In this way, it is not entirely clear how Hume’s skeptical philosophy could function as the medicine he thinks it is. In fact, the problem here is two-fold. In the first instance, as I have already suggested, it is not clear how the skeptical doubts his philosophy inspires could prevent the mind, once recovered from the gloomiest depths, from again pretending to soar up to lofty subjects in fact inaccessible to human understanding. Hume tells us that “reason must remain restless, and unquiet, even with regard to that scepticism, to which she is driven by. . . seeming absurdities and contradictions.”155 Again, Hume says, “’tis almost impossible for the mind of man to rest, like those of the beasts, in that narrow circle of objects, which are the subject of daily conversation and action.”156 Admitting that even he himself is not immune to this drive, Hume says, “it wou’d necessarily happen, that from my very weakness I must be led into such enquiries” as “speculations without the sphere of common life.”157 And that suggests that even upon the experience of and recovery from philosophic melancholy, a return to such pretensions is inevitable for creatures like us. Second, and perhaps worse, it is not clear how the skeptical philosophy, insofar as it is kept in mind, could leave room for belief even in the sphere of common life. The considerations

153EHU 12.25, SBN 162. 154See T 1.1.5, 1.3.6; EHU 4–5. 155EHU 12.20, SBN 157–58. 156T 1.4.7.13, SBN 271–72. 157T 1.4.7.13, SBN 272.

92 which led Hume into his melancholy have shown human belief and reasoning to be based on a principle or “quality” of our imagination, “by which the mind enlivens some ideas beyond others (which seemingly is so trivial, and so little founded on reason).”158 It has shown, moreover, that this principle leads us to “embrace. . . manifest contradiction”159 and reveals the things most “curiously enquir’d after” to be things about which we can have no “hope of ever attaining satisfaction.”160 Amongst these unattainable curiosities are “the causes of every phænomenon” or “operating principles” that we ordinarily take to “reside in the external object.”161 Many of such causes would seem to be among the mundane topics of daily conversation. But according to the skeptical results of Hume’s science of man, when we speak of them “we either contradict ourselves, or talk without meaning.”162 How then should we form belief even in the most familiar and ordinary things, when we hold Hume’s skeptical conclusions in view? It seems Hume’s discussions in 1.4.7 show that only natural forces, encouraged by counterbalancing temperaments and activities, could restore even our “indolent belief in the general maxims of the world.”163 Hume may be comfortable with the idea that balanced, healthy judgment of the sort he hoped a dose of skeptical philosophy could help inculcate, while perhaps aided to a degree by reasoning upon experience, still requires the guidance of a benevolent nature we cannot understand but nonetheless ought to submit to. But this reliance on nature puts us in an awkward position, in which we must put trust in our natural tendencies and inclinations, and hope that nature can steer us between the Scylla of extravagant fictions and the Charybdis of skeptical oblivion. When asking ourselves “how far we ought to yield to the illusions” of our imagination, Hume reports, we find that “this question is very difficult, and reduces us to a very dangerous dilemma, which-ever way we answer. . . We have. . . no choice left but betwixt a false reason and none at all.”164 It seems the only response we can muster is, in each case of reasoning, to exercise our judgment according to our inclinations. But, again, why should our inclinations keep our thought safely within the mortal realm, when it is

158T 1.4.7.3, SBN 265. 159T 1.4.7.4, SBN 266. 160T 1.4.7.5, SBN 266–67. 161T 1.4.7.5, SBN 266–67. 162T 1.4.7.5, SBN 267. 163T 1.4.7.10, SBN 269. 164T 1.4.7.6–7, SBN 267–68.

93 so natural to go beyond it? And why should they save us from melancholic gloom, when this sort of excess, too, is natural to those of a philosophic temperament? What use is the experience of skeptical philosophy—what medicine—if we must in the end rely on our natural robustness to find a middle path? It seems to be at best—to borrow a phrase from Hume’s own criticisms of fruitless philosophical pursuits—“a palliative remedy,” perhaps helpful for a while, but unable to remove the root of the infirmity.165 Like Hume, Kant thought that philosophy can, or, indeed, “must also act (therapeuti- cally) as a medicine.”166 And, also like Hume, Kant saw parts of his philosophy as addressing “that temperament and our various talents (such as imagination and wit) which incline to allow themselves a free and unlimited activity” and “are in many ways in need of discipline.” Accordingly, Kant thought that part of philosophy’s utility is “to guard. . . against extrav- agance and error” typical of the overly sanguine.167 As I will argue in the next chapter, he viewed Hume’s censorship of the pretenses of past metaphysicians as acting with this humane aim. Indeed, Kant praised Hume,168 and openly adopted from Hume what he saw to be a salutary lesson. He went so far as to suggest that “Hume’s principle, not to drive the use of reason dogmatically beyond the field of all possible experience” was “the result of the entire Critique,” Kant’s seminal text.169 But Kant remained critical of Hume’s medicine for leaving us with little guidance about its proper usage, which he thought must always rely on haphazard judgment: According to Kant, applying Hume’s principle was a task for “which we are advised to determine for ourselves as it were mechanically (something from one side, and something from the other), and by which no one is taught any better.” This characterization seems apt, insofar as healthy belief depends for Hume upon a balance between temperaments, which incline us to one side or the other. But this procedure makes us reliant on nature, or natural inclinations; it gives no clear insight into the exact boundaries of our knowledge. Insofar as our nature is fickle, both dogmatism and skepticism remain apt to return. Kant’s iteration on Hume’s

165See T 1.4.2.46, SBN 211–12. 166TeFP, 8:415. Kant makes a similar point when he says that “philosophers. . . prescribe the diet of the mind” (VKK 2:271). 167A710–11/B738–39. 168See, especially, §4.3 below. 169P 4:360–61.

94 medicine—the ‘criticism’ developed in his Critique of Pure Reason—is instead to proceed “according to principles” which determine both the medicine’s and reason’s proper use.170 Such principles, Kant promised, allow us to reach a middle path without relying on blind instinct or mechanical balancing procedures. They would do this first by drawing a boundary beyond which we can have no insight, crushing dogmatism. But the very same principles would allow us to overcome exactly the kinds of charges that Hume brought against reason and the imagination in 1.4.7: They would ground our knowledge and reasoning in the valid use of the understanding, remove the contradictions which appear endemic to reason, and restore those most curiously “enquir’d”-after causes to their rightful place as intelligible features of our ordinary empirical knowledge.171 In this way, Kant means to offer a cure for skepticism, and not just its moderation. Let us then turn to Kant, to see how his response to skepticism can go beyond that of moderation, and provide benefits beyond inculcating a degree of humility—including those I described in §2.7. Such a response can explain our ordinary insight into the world through experience, and indeed do so in a way which a skeptic can embrace.

170P 4:360. 171See the Critique of Pure Reason’s System of all Principles of Pure Understanding (A148–235/B187– 294), The Antinomy of Pure Reason (A405–567/B432–595), and Analytic of Concepts (A65–130/B90–169), respectively.

95 4.0 KANT’S OFFER TO THE SKEPTICAL EMPIRICIST

4.1 A NEW QUESTION

Kant viewed Hume as a skeptical empiricist. Kant portrayed skeptics first and foremost as wary of our pretensions to engage in metaphysics, an alleged science in which we attempt to extend our knowledge a priori—that is, in advance of experience. Kant portrayed empiricists as committed to a view of the mind on which our concepts, and all substantive knowledge in which they figure, derive from experience. In viewing Hume as a skeptical empiricist, then, Kant portrayed him as challenging our pretensions to metaphysics in a particular way—namely, by challenging the status of concepts which Kant took to be pure concepts, or ‘categories.’ Pure concepts originate in the understanding prior to experience, and for that reason seem amenable to use in metaphysics. Although skeptical empiricists did not use the term ‘pure concept,’ their views had the effect of challenging the status of those concepts as pure. Kant thought that revealing those concepts’ allegedly pure status to be illusory would deprive us of their use in metaphysics, as there can be no a priori use of concepts that are found to derive from experience.1 According to Kant, Hume gave an empiricist account of a prime candidate for a pure concept—namely, the concept ‘cause.’2 On that account,“the concept of a cause” is “an illusion” originating in a “custom” of associating certain events with others that one has experienced regularly following them,3 and so the concept ‘cause’ ultimately derives from experience. The success of such an account would provide grounds for doubting any alleged metaphysical knowledge involving the concept ‘cause’—for example, “the proposition, that the world must have a first beginning”4—as it would be revealed as an attempt to apply the

1I provide textual evidence for these readings below, primarily in §§4.2–4.3. For a note on abbreviations, editions, and citation styles used for Hume and Kant’s writings, see my “Note on Citations of Hume and Kant” below, beginning on p. 160. 2That such a concept is pure, Kant thinks, is clear from the fact that it “manifestly contains the concept of a necessity of connection with an eﬀect and of the strict universality of the rule” (B5), given that “experience never confers on its judgments true or strict. . . universality” (B3–4; cf. A734/B762). 3KpV 5:51. 4B18.

96 materials of experience beyond all experience. Kant thought Hume’s account raised a doubt that could be generalized to all ‘pure’ concepts,5 and so be used to take down metaphysics in one fell swoop.6 In doing so, however, it would call into question any extension of our knowledge which does not draw on actual experience—or, as Kant put it, all synthetic a priori knowledge. That, for Kant, would not only include the metaphysician’s supposed knowledge of God, freedom, and immortality, but also the actual systems of knowledge of mathematics and general natural science, as well as the principles we necessarily employ in experience—including, for example, the principle that everything that happens has a cause.7 Calling these latter principles into question would in turn call into question even ordinary empirical knowledge, which employs those principles. In this way, Kant concludes, “Hume’s empiricism in principles leads unavoidably to skepticism” and, indeed, “a universal skepticism.”8 There are countless questions to ask about Kant’s relationship to skeptics, to empiricists, and to Hume, and a variety of such questions have long animated scholarship on Kant’s theoretical philosophy. In this chapter, I will address the following questions: Did Kant expect skeptical empiricists, including Hume, to accept the critical philosophy Kant develops

5B19–20; P 4:260–61. 6P 4:258. 7On the conflict of skeptical empiricism with mathematics and general natural science, see B127–28, A760/B78, A765/B793; KpV 5:13, 5:51. On the conflict of skeptical empiricism with the principles of experience, including the causal principle, see A760/B788, A765/B793, KpV 5:13. 8KpV 5:52; cf. A760/B789. Kant clarifies that Hume himself did not see his skepticism as extending even to mathematics. But Kant thought this was only because Hume did not realize that mathematical knowledge was synthetic (KpV 5:52). The accuracy of Kant’s reading here need not concern us. Of greater interest is Kant’s description of skepticism in general as expanding even to empirical knowledge:

Skepticism originally arose from metaphysics and its unpoliced dialectic. At ﬁrst this skepticism wanted, solely for the beneﬁt of the use of reason, to portray everything that surpasses this use as empty and deceitful; but gradually as it came to be noticed that it was the very same a priori principles which are employed in experience that, unnoticed, had led still further than experience reaches—and had done so, as it seemed, with the very same right—then even the principles of experience began to be doubted. (P 4:351)

This passage suggests that it is a tendency of skepticism in general for its doubt to go beyond metaphysics and attach even to empirical knowledge, insofar as the possibility of the latter depends on the same a priori principles which are illicitly used in metaphysics. Skeptical empiricism leads to a distinctive kind of doubt about our pretensions to extend our knowledge a priori—namely, by focusing on the origin of our concepts. But expansion beyond the topics of traditional metaphysics—God, freedom, and immortality—is not necessarily unique to skeptical empiricism. On Kant’s view, skepticism in general tends to expand even to empirical knowledge.

97 in his Critique of Pure Reason? If so, how, or why, or under what conditions did Kant think the skeptic would accept his conclusions? The first of these questions is not the same as the question whether Kant intended his arguments in the Critique to refute a skeptical empricist, or to answer Hume, in a way that does not beg the question against their empiricism. In recent years, that question has become perhaps somewhat tired, despite the continued lack of consensus as to its proper answer. More importantly, I think the question so-framed has been an obstacle to understanding Kant’s true intentions. This is because its framing easily leads us to overlook the possibility that Kant intended to change the skeptical empiricist’s mind in a way that, in a certain sense, does beg the question against her views—or, at very least, does not constrain itself to using only resources that a skeptical empiricist can embrace while remaining committed to her empiricism. Indeed, as I will argue below, Kant intends to change the skeptic’s mind not by arguing from such limited resources to the falsity or incoherence of skepticism or empiricism, but instead by developing and offering an alternative conception of the mind’s relation to the objects of knowledge. Though this alternative is incompatible with empiricism, even fundamentally opposed to it, Kant believes that skeptical empiricists can willingly embrace his alternative, and in doing so cease to be both skeptics and empiricists. He thinks his alternative can appeal, first, because it satisfies explanatory aims which skeptics necessarily, if sometimes only latently, harbor, and, second, because skeptics tend to feel dissatisfaction and doubt about their own skeptical position, both of which allow them to dissociate from that position sufficiently to entertain alternatives. Kant’s method for changing the skeptical empiricist’s mind, then, builds upon his view about an instability inherent to skepticism—a view that in turn explains how skeptics can be reached by resources which extend beyond their empiricism. In what follows, I will argue that Kant develops these resources throughout his Tran- scendental Analytic, and explains how they can appeal to the skeptic largely in the often overlooked Discipline of Pure Reason. Before doing this, though, I will introduce, in §4.2, two dominant readings of Kant’s relationship to Hume, and to skepticism in general. The first views the task of Kant’s Analytic to be refuting the skepticism which Kant found crystallized in Hume’s discussions of causation, while the second emphasizes Kant’s aim to appropriate,

98 rather than reject, Hume’s censorship of reason’s pretensions. Introducing these readings will help to explain in more detail what Kant thought the skepticism he associated with Hume and empiricism was.9 I then explain that these two readings have appeared mutually exclusive due to a narrow conception of the options for responding to skepticism, encouraged by a traditional reading of the form of the Analytic’s central arguments. Rejecting this conception, I then argue, creates space for the reading I want to advance. In §4.3, I reconsider Kant’s attitude toward Hume and skepticism in light of his most explicit remarks about them. These remarks, which include both compliments and concerns, suggest that Kant views Hume and skeptics more as allies than as opponents, and diagnoses their excessive censorship of reason and the understanding as overlooking a possible explanation of the kind of knowledge which they too hastily reject. This suggests that skeptical empiricists could be encouraged to progress beyond their skepticism, if provided with the explanation that

9In his critical writings, Kant usually speaks of ‘skepticism’ or ‘doubt’ in connection with Hume or empiricism. See Aix–x, B19–20, B22–23, B127–28, A758–69/B786–97; P 4:256–64, 4:272, 4:310–13; KpV 5:12–14, 5:50–57. In the body text, I will only discuss Kant’s attitudes toward a kind of skepticism he associates with Hume and empiricism. Various interpreters have read Kant as responding to this kind of skepticism, often by responding to Hume. See Bennett(1966), 100–02; Wilkerson(1971), 351–52; Rorty(1979), 77; Wolff(1981), 28; Hoppe(1983), 61–62; Guyer(1987), 67; Henrich(1989), 36–38; Guyer(2003), 8–10; Stern (2006), 112; Forster(2010), 40–41; Chance(2011), 342; Chance(2013), 225–26, 234–36; Sommerlatte(2016), 448–50. Against this view, see Brook(1997), 96; Engstrom(1994), 370–76; Anderson(1998), 146; Hatfield (2001), 188–89; Ameriks(2003), 64; Hatfield(2003), 178–84; Watkins(2004), 453, 486–87; Allison(2008b), 538; Dyck(2011), 489–96. I discuss some of these interpretations in §4.2 below. Kant has also been taken to respond to other kinds of skepticism—for example, a ‘Cartesian’ or ‘veil of perception’ skepticism, which doubts that thought can transcend an inner realm in a way that is deemed necessary for knowledge of objects outside of that realm. Kant tends to discuss this view under the label of “problematic” or “material idealism” rather than “scepticism,” and only a few times associates it with ‘doubt’ (Bxxxix, B124–25, A377, B519n). In the second half of the twentieth century, many Anglophone interpreters viewed Kant’s Analytic, and especially his Transcendental Deduction, as primarily concerned with this kind of skepticism. See Strawson(1966/2002), 88, 97; Stroud(1968/2000), 9–10; Rorty(1970), 207; McCann(1985), 71; Brueckner(1984), 197; and Guyer(1987), 67. Against this view, see Ameriks(1978), 273; Allison(1983/2004), 159–60; Lear(1984), 220ff; Engstrom(1994), 360–70; Guyer(2003), 5–6; Forster (2010), 6–12; and Dyck(2011), 446–81, 489–96, esp. 495. Some interpreters now view Kant’s Critique as responding to a ‘Pyrrhonian skeptic,’ who opposes equally specious claims of dogmatic metaphysics in order to bring about doubt. See Guyer(2003), 4–10, 17–18; Forster(2010), 44ff; Stern(2006), 104ff; and Stern(2008), 273ff. Though Kant never explicitly refers to Pyrrhonian skepticism in the first Critique, he does describe a “sceptical method” which opposes specious claims, and indeed often praises this method. In contrast to “scepticism—a principle of technical and scientific ignorance, which undermines the foundations of all knowledge and strives in all possible ways to destroy its reliability and steadfastness,” the skeptical method “aims at certainty,” “seeks to discover the point of misunderstanding in the case of disputes” (A424/B451–52), and is “favoured” by “the transcendental dialectic” (A507/B535). The same favorable tone appears in Kant’s descriptions of Pyrrhonian philosophy (though not certain “dogmatic” followers of Pyrrho) in lectures (see BL, 24:208–10). I find it hard then to see Kant as defending against Pyrrhonian skepticism. For a similar criticism of an anti-Pyrrhonian reading of Kant, see Chignell and McLear(2010), 233–34, 237–38.

99 they overlook. In §4.4, I provide textual evidence that Kant intends to provide the core of this explanation in the Transcendental Deduction of the Pure Concepts.10 On my reading, the principal aim of the Deduction is to provide an explanation of an overlooked possibility, rather than a proof. I call this the explanatory reading of the Deduction. In §4.5, I show that this reading is consistent with the Deduction’s content, and helps to resolve a puzzle about its structure. In the remainder of the chapter, I respond to the worry that the skeptical empiricist would fail to be moved by Kant’s explanation. In §4.6, I explain that Kant believes his explanation can appeal to a basic human drive to answer questions that go beyond our experience—a drive which the skeptic cannot succeed at removing. It may seem that empiricism provides grounds for quieting this drive. But, according to Kant, empirical grounds can never suffice to prove that we cannot know things independently of experience. On the contrary, the insufficiency of those grounds leaves room for the skeptical empiricist to harbor hope for a priori knowledge. The explanation provided by Kant’s Deduction can then appeal to the skeptic’s lingering hope because it explains how we can know things a priori. Though ultimately Kant thinks we cannot have metaphysical knowledge of the kind most pined after, his explanation still renders intelligible our a priori knowledge of the principles employed in experience, such as that everything which happens has a cause. Kant’s explanation of our knowledge of such principles in turn explains the possibility of empirical knowledge which employs those principles. If this is right, the Deduction’s explanation can both appeal to the skeptic and cure her skepticism. In §4.7, I acknowledge that various moods which may accompany skepticism can stand in the way of this appeal. But I argue that we need not see this as reflecting a flaw in Kant’s offer of a cure. What emerges throughout is a picture on which Kant’s response to the skeptic is both more amicable and more sound than it is often thought to be. By explaining the possibility of our synthetic a priori knowledge, Kant’s Transcendental Deduction offers a way out of skeptical empiricism, without compelling it.

10I use ‘Deduction’ (capitalized) to refer to this argument, and ‘deduction’ (lowercase) to refer to deductions in general.

100 4.2 TRADITIONAL READINGS OF KANT ON HUME

In arguing for my reading, I will partly agree with and partly reject two dominant interpretations of Kant’s relationship to Hume, both about as old as the Critique itself. The first interpretation, perhaps more familiar to an anglophone audience, is one on which the Critique’s central aim is to respond to a challenge Kant found in Hume’s philosophy. I call this the oppositional reading. A second interpretation is perhaps less familiar now, though enjoying a recent resurgence. This is the view that Kant was not so much defending against Hume as adopting and completing Hume’s project of censoring metaphysics. I call this the inheritor reading.11 On the oppositional reading, Kant’s aim in the Critique is primarily anti-skeptical. Ac- cording to this reading, a major goal of that text, even the organizing theme at least of its Transcendental Analytic, is to respond to the skeptical worries which he associates with Hume, about the valid application of concepts like ‘cause’ and ‘substance’ to the objects of experience—and to do so in a way that could change the skeptic’s mind, without begging the question against her. As Paul Guyer puts it, “Kant’s purpose is to attack skepticism” and “to undermine a position which may as well be called empiricism;”12 “refuting Humean scepticism. . . is the project. . . of the first half of the Critique.”13 Dieter Henrich writes that Kant criticizes Locke, Leibniz, and “physiologists of reason” for failing to do “what ultimately matters in philosophy: justifying the claims of reason against skepticism”—the task of critical philosophy.14 As Gary Hatfield describes the view (though without endorsing it): “The core of the standard view is that Kant wanted to justify the categories in the face of skepticism. The Deduction was to ‘defeat the skeptic.’”15

11As noted in Chance(2013), 213–14, within just one year after the publication of the Critique’s ﬁrst edition, Kant had already been portrayed in print as “disputing” or “rejecting [bestreiten]” Hume’s philosophy. This characterization appeared in an anonymous translator’s preface to the 1782 German edition of Reid’s Inquiry into the Human Mind. In contrast, even while the Critique was still in press, some of Kant’s contemporaries, such as Johann Georg Hamann, had already come to view Kant as continuing Hume’s attacks on the pretensions of reason. Hamann thus referred to Kant in letters as a “Prussian Hume.” See, for example, a 1781 letter to Herder in Hamann(1955), 4:293, 298, 343. 12Guyer(1987), 67. 13Guyer(2003), 9. 14Henrich(1989), 36–37. 15Hatﬁeld(2003), 178.

101 The oppositional reading views the Transcendental Deduction as the lynchpin in this anti-skeptical project, defending against Hume’s challenge to the rights of reason by arguments in defense of those rights. When introducing the Transcendental Deduction of the Pure Concepts—an “enquiry” most “important for exploring the faculty which we entitle understanding, and for determining the rules and limits of its employment,”16 Kant portrays that Deduction as answering a “question of right (quid juris)” about our use of such concepts, also called ‘the categories.’17 Unlike “many empirical concepts” for which “experience is available for the proof of their objective validity,”18 the categories, as ‘pure’ concepts, are not “grounded in experience” and so “arouse suspicion. . . in regard to their objective validity and limits of their own employment.”19 Thus Kant ﬁnds a “transcendental deduction of [their employment] necessary.”20 Proponents of the oppositional reading tend to view an empiricist such as Hume as the source of this quid juris. Henrich articulates this picture of “the skeptic—here in the guise of the empiricist” when he writes:

A legal dispute originates when a party’s claim has been challenged by an opponent, so that a court case must be opened. This happened in philosophy when the skeptic challenged the claim of reason to be in possession of a priori knowledge of objects. The dispute makes indispensable an investigation into the origins of such knowledge. To the extent to which a deduction can be produced, the claim of reason becomes deﬁnitely justiﬁed and the challenge of the skeptic is rejected. This is the aim of the Transcendental Analytic.21

Henrich’s view of the challenge as arising from a skeptical empiricist is naturally applied to Hume. After all, Hume had argued in his Enquiry Concerning Human Understanding that all of the ideas we employ in our thought (so, for Kant, our concepts) trace back to “the materials aﬀorded to us by the senses and experience.”22 “[I]f a proper use were made of this ﬁnding,” Hume thought, we could rid ourselves of unrightful pseudo-ideas, and in particular,

banish all that jargon, which so long has taken possession of metaphysical reasonings and drawn disgrace upon them. When we entertain, therefore, any suspicion, that a philosoph-

16Axvi. 17A84/B116. 18A84/B116. 19A88/B120. 20A88/B121. 21Henrich(1989), 38. 22EHU 2.5, SBN 19.

102 ical term is employed without any meaning or idea (as is but too frequent), we need but inquire, from what impression is that supposed idea derived? And if it be impossible to assign any, this will serve to conﬁrm our suspicion. By bringing ideas into so clear a light, we may reasonably hope to remove all dispute, which may arise, concerning their nature and reality.23

We can call this procedure of searching for a sense impression which gives rise to the idea at issue Hume’s test. Given that the categories have their “birthplace” in the understanding alone, abstracting from our sensibility, and so are prior to experience,24 they seem to be prime candidates to arouse an empiricist’s suspicion and be written off as unrightful in virtue of failing Hume’s test. Hume does not think the concept ‘cause’ is meaningless. Indeed, he tries to show that an idea of the ‘necessary connexion’ which we believe obtains between a given cause and its effect passes his test. So he does not conclude that any talk of one thing’s causing another must lack meaning altogether. But he does find that it must mean something different, or less, than we ordinarily take it to mean. He comes to this conclusion when he finds that our idea of ‘necessary connexion’ arises from an impression we feel of our own mind’s being determined, out of habit, to expect a given effect to follow when perceiving or considering a given cause.25 Hume concludes from this discovery about the origin of the idea that “we have really no idea of a power or efficacy in any object, or of any real connexion between causes and effects.”26 We can still intelligibly and clearly speak of one object’s causing another, if all we mean is that objects like the former regularly precede objects like the latter.27 But regular succession is not necessity. Since, for Hume, our idea of necessity arises from our feeling a habitual determination of our own mind, we “confound ideas” if we take this idea to be that of a real connection between objects.28 Kant interprets the

23EHU 2.9, SBN 21–22; cf. 12.34, SBN 165. 24A65/B90; A86/B119. 25See T 1.3.14.1, SBN 156; EHU 7.28, SBN 75–76; 7.30, SBN 78–79. 26T 1.3.14.27, SBN 168, my emphases. See also: “The efficacy or energy of causes is neither plac’d in the causes themselves, nor in the deity. . . but belongs entirely to the soul, which considers the union of two or more objects in all past instances. ’Tis here that the real power of causes is plac’d, along with their connexion and necessity” (T 1.3.14.23, SBN 166). 27As Hume puts it, there is a “natural” relationship between the objects we call ‘causes’ and ‘effects’— namely, that one object regularly precedes the other. See the first of Hume’s two proposed definitions of cause at T 1.3.14.35, SBN 172; 1.3.14.31, SBN 169–70; and EHU 7.29, SBN 76–77. 28T 1.3.14.27, SBN 168; cf. T 1.4.7.5, SBN 266–67. This holds, for Hume, because he thinks that all our ideas are either copied directly from impressions, or are composed of various ideas which are copied directly

103 result of this “empirical derivation” of Hume’s as showing that the concept ‘cause’ has a merely “subjective necessity.” If the derivation is correct, the application of such a concept to the objects of experience “passes off the resulting subjective necessity (i.e., habit) for an objective necessity (from insight).”29 This image of counterfeit substitution can create a different kind of suspicion about our right to say that one thing causes another or that the latter follows necessarily after the former. Our concept of cause can then seem “fraudulent and deceptive,”30 rather than meaningless. Kant thinks the concept is neither meaningless nor deceptive. Insofar as his Deduction of the categories is a deduction of pure concepts, it clearly makes no attempt to show that the categories pass Hume’s test. Indeed, Kant tells us that a story about how such “universal concepts” derive from “particular perceptions” in experience—a “line of inquiry” Kant associates with Locke’s empiricism—does not amount to a deduction. “For in view of their subsequent employment, which has to be entirely independent of experience, they must be in a position to show a certificate of birth quite other than that of descent from experience.”31 But even if Kant has no intention to show that the categories pass Hume’s test, he can still seem interested in quieting an empiricist’s suspicion with his Deduction, given its argumentative form. Though the two editions’ versions of this “proof”32 differ substantially, they are both preceded by the same sketch. Here, Kant says: “The objective validity of the categories as a priori concepts rests. . . on the fact that. . . through them alone does experience become possible.”33 If it can be made plausible that the application of the categories to objects of knowledge is a condition on the possibility of experience, then the argument would seem to have overwhelming force. As Ralph Walker puts it: “Kant intended his transcendental arguments to be arguments against the sceptic. They are supposed to work by showing that there is an incoherence in the sceptical position: the fact that there is experience, or knowl- from impressions. The idea of ‘necessary connexion’ is, for Hume, of the former kind. So it must be an idea of the same object which the impression it is copied from is an impression of. And, according to Hume, that object turns out to be something in the mind. See T 1.1.1.7, SBN 4; T 1.3.14.4, SBN 157; EHU 7.4, SBN 62. 29P 4:258; cf. B5; B127–28; P 4:312, KpV 5:13. 30KpV 5:51. 31A86–87/B119. 32A84/B116; B145. The German here is ‘Beweis.’ 33A93/B126.

104 edge, renders that position untenable.”34 Acquiescence to the arguments would seem to be inevitable, since it seems that no one could deny that experience is possible. Robert Wolff writes that “Since not even Hume is prepared to deny that he is conscious,. . . Kant will, if he can make his argument, have answered Hume.”35 Presumably, even skeptical empiricists would be compelled to accept the objective validity of the categories, and recognize that their rightful use is presupposed rather than questioned by any test of concepts which looks to experience for validation. Such recognition would presumably move the empiricist to give up her empiricism, and, in the same breath, remove her skepticism about philosophical principles which employ those pure concepts. If all this is correct, Kant’s “refutation of” or “answer to” Hume would constitute a cure for Hume’s skepticism, even if offered in a spirit of opposition. Kant’s explicit invocations of Hume when he discusses the Critique in other texts also seem to encourage reading Kant as intending the Deduction, and indeed the Critique as a whole, as responding to a challenge he found in Hume’s writings. In the preface to the Pro- legomena, for instance, Kant portrays Hume’s discussion of ‘cause,’ one of Kant’s categories, as generating a quid juris. “Hume,” he says, “started mainly from a single but important concept in metaphysics, namely, that of the connection of cause and effect. . . , and called upon reason, which pretends to have generated this concept in her womb, to give him an account by what right she thinks.”36 Kant says that this “warning [Erinnerung]37 of David Hume was the very thing that many years ago first interrupted my dogmatic slumbers” and led Kant to conceive his critical project.38 Again, in the Critique of Practical Reason, Kant reports that “David Hume. . . can be said to have really begun the assaults on the rights of pure reason which made a thorough investigation of them necessary.”39 Kant repeatedly insists that Hume’s analysis of ‘cause’ as arising through habit or custom makes the concept

34Walker(1999), 13. Walker notes that despite inductive skepticism’s association with Hume, and Kant’s concern with skepticism concerning the application of the concept ‘cause,’ Kant is also concerned with a ‘global’ skepticism (13n1). Whether Walker considers Hume to be a such a global skeptic is left unclear. 35Wolﬀ(1981), 28. 36P 4:257. 37Hatﬁeld has “remembrance.” Either translation would be natural and consistent with Kant’s usage of this term elsewhere. See 10n9. 38P 4:260. Cf. A757/B785; TeFP 8:14; FM 20:319. 39KpV 5:50.

105 out to be “fraudulent and deceptive,” an “illusion,” and a “bastard of the imagination.”40 His insistence seems to indicate an interest in rescuing ‘cause’ and certain other concepts central to metaphysics from a threat. Kant’s expresses dissatisfaction with Hume’s “empirical derivation” of the concept ‘cause’ as “aris[ing] from repeated association” just before the start of the Deduction,41 again reinforcing this reading of Kant’s aims in the Deduc- tion. Continued attention to these passages, and to Kant’s discussion of a causal principle in the Second Analogy, has reinforced this interpretive tradition. On the oppositional reading, T.E. Wilkerson writes, “It is perfectly natural to regard Hume as Kant’s unseen opponent throughout the argument of the Transcendental Analytic, and natural, therefore, to regard the Second Analogy as the most important later stage of that argument.”42 Various parts of Kant’s corpus, then, seem to support an oppositional reading, on which the Critique is primarily anti-skeptical. This oppositional reading is far from universally accepted. A second, contrasting interpretive tradition views Kant as having little or no concern with convincing Hume or his followers to abandon skepticism or empiricism. Rather, on this view, Kant sees Hume primarily as a predecessor and an inspiration. He sees himself as inheriting and continuing Hume’s project of curbing reason’s pretensions, albeit on better grounds. I call this the inheritor reading. Eric Watkins helpfully expresses this reading: “Hume is important to Kant not because Hume’s position stands in need of refutation, but rather because Hume develops provocative critical insights into a variety of particularly fundamental arguments and issues, insights that Kant wanted to accommodate within his own system.”43 Some of these insights presumably included the illegitimacy of applying concepts like ‘cause’ to the traditional objects of metaphysics, especially to God. As Karin de Boer puts it, “Kant was ﬁrst and foremost struck by Hume’s account of causality because he considered it to undermine proofs for the existence of God and, more generally, the putative synthetic a priori knowledge to which metaphysics aspired.”44 A sense of Kant’s aﬃnity with Hume, especially with regard to

40P 4:257–58; KpV 5:51. 41B127–28. 42Wilkerson(1971), 351–52. 43Watkins(2004), 486–87. 44de Boer(2019), 377.

106 their negative appraisal of speculative theology, was already in the air by the time the first edition of the Critique went to press, when some of Kant’s contemporaries referred to him as a “Prussian Hume.”45 And it remains an appealing reading of Kant’s attitude toward Hume in both editions of the Critique and in the Prolegomena. Hatfield writes: “In none of the three works was Kant’s main aim to ‘answer the skeptic.’ His primary aim was what he said it was: to firmly establish the boundaries of metaphysics, by discovering the elements of human cognition and fixing its proper domain.”46 Hence, Karl Schafer concludes that “Kant and his contemporaries were right to see Kant as aiming more to complete Hume’s project than to refute it.”47 As Hatfield points out, Kant introduces his task in the Critique as that of “decid[ing] as to the possibility or impossibility of metaphysics in general, and determin[ing] its sources, its extent, and its limits.”48 His finding the limits to be wider than the skeptic claims can seem to support the oppositional reading. But insofar as those limits are much narrower than the Wolffian metaphysicians of Kant’s day presumed (if indeed they concerned themselves with such limits at all), Kant’s interest in determining the limits suggests that he is closer to embracing than to rejecting Hume. And that seems consistent with Kant’s own portrayal of his project in the Critique, whose “utility,” he says, “ought properly to be only negative, not to extend, but only to clarify our reason, and keep it free from errors.”49 He even goes so far as to portray his findings concerning those limits as expressible in terms of what he calls “Hume’s principle[:] not to drive the use of reason dogmatically beyond the field

45See note 11. 46Hatfield(2001), 189. Hatfield notes that the increased number of discussions of Hume in the Prole- gomena and second edition of the Critique, compared to the first edition, was motivated by Kant’s desire to “distinguish his position more sharply from Hume’s” in the face of misinterpretations of the first edition by Kant’s contemporaries (188–89). While this may be correct, it is worth noting that Kant’s longest and clearest statement of the differences between his critique and Hume’s censorship appears in the Discipline of Pure Reason, and that stretch of text remains completely unchanged between the two editions. There, Kant already complains that Hume overextends his censorship of metaphysics to the principles we employ in experience, and diagnoses what he takes to be the causes of this overextension. The presence of this diagnosis in first edition puts pressure on Hatfield’s claim that “Kant certainly had no thought, prior to the initial responses to the Prolegomena and Critique, that ordinary experience could use philosophical vindication” (207). A diagnosis which shows how subtly motivated doubt about the principles employed in experience goes wrong just is one way to give a philosophical vindication of experience. For accounts of Kant’s diagnosis, see §4.3 below, as well as Engstrom(1994), 371–75; Chance(2011), 333–39; Schafer(forthcoming), §§4–5. 47Schafer(forthcoming), §3. 48Axii. 49A11/B25.

107 of all possible experience.” This principle, he continues, can be thought to be “the result of the entire Critique.”50 This, again, can seem to show that he saw no need to overturn Hume’s philosophy, whose core he accepted. Making this point, Manfred Kuehn articulates the inheritor reading clearly:

Kant and Hume aim at the very same thing (or, at least,. . . Kant believed he was aiming at the very same thing as was Hume), namely the determination of the limits of metaphysics and human knowledge in general. Kant clearly thought that he was the executor of Hume’s philosophical will. This means, however, that Kant was not primarily concerned with ‘answering’ Hume or refuting skepticism. His critical philosophy is in a fundamental sense a justiﬁcation of Hume’s principle.51

If the Critique’s aims are only negative, it seems it need not attempt to convince a skeptic that we have synthetic a priori knowledge. That Kant did not try to convince someone like Hume of his positive doctrines also seems to be supported by Kant’s insistence that Hume’s own aim was really just to censor metaphysics. Hume “never put into doubt,” he says, whether “the concept of cause is right, useful, and, with respect to all cognition of nature, indispensable.” Rather, what Hume questioned was whether such a concept “is thought through reason a priori, and in this way has an inner truth independent of all experience, and hence also a much more widely extended use that is not limited merely to objects of experience.”52 Kant says that Hume adopts his account of “the concept of cause, so as to deny to reason any judgment of God, freedom, and immortality.”53 Again, his “‘attacks are directed,” Kant says, “chieﬂy against” “the dialectical pretensions of reason.”54 Though Kant acknowledged the doubt empiricism casts upon mathematics, general natural science, and empirical cognition, he can seem to treat this as an unfortunate side eﬀect of empiricism and not a proper threat to those domains

50P 4:360–61. 51Kuehn(1983), 191. 52P 4:258–59. It may appear surprising or misleading of Kant to portray Hume as questioning or even so much as contemplating that concepts like ‘cause’ could be “thought through reason a priori.” But Hume was surely aware that past philosophers and theologians, especially in mainland Europe, portrayed themselves as learning substantive truths about the world or God by thinking certain concepts ‘through reason a priori.’ Kant would be right, then, in thinking that Hume was concerned with disabusing such ﬁgures of their pretensions, even if he thought that empiricist doctrines exclude the very possibility of extending our knowledge this way. For Hume, this exclusion was part of the appeal of empiricism. I return to this point in §4.6. 53KpV 5:13. 54A768/B796. Cf. A745/B773.

108 of knowledge, given that he seeks to constrain metaphysics through a different means. The inheritor reading also gains traction through considerations of charity, given that the oppositional reading can seem to cast a poor light on Kant. As interpreters of both camps have noted, Kant’s major arguments can seem to beg the question against a skeptical empiricist like Hume. It is not immediately obvious, for example, how he could get an empiricist to agree that the valid application of the categories is required for experience, insofar as the empiricist nurtures grounds for doubting that any such concept could figure in substantive knowledge independently of experience.55 Alternatively, if Kant could make the categories’ valid application plausible, it may then seem he is working with a more robust conception of experience, or empirical knowledge, than the skeptical empiricist would grant.56 Either way, Kant seems to assume things the skeptic would deny, and similar points have been made about some of Kant’s other candidate anti-skeptical arguments.57 This allegedly blatant failure has motivated some readers to reject the claim that Kant aims to refute skepticism with such arguments in the first place. Hence, Hatfield criticizes the “standard” anti-skeptical reading of his aims for distorting contemporary assessments of Kant’s success. On that oppositional reading, “[t]he Transcendental Deduction was to ‘defeat the skeptic,’ and one of the most frequent criticisms in recent decades has been that it failed to do so because it argued inconsequently, perhaps by begging the question. This (mistaken) construal of the task of the Deduction has misled interpreters for decades.”58 If Kant’s arguments really do beg the question against Hume, charity might then call for a reinterpretation of Kant’s relationship to Hume and his skepticism. And the considerable textual evidence can reinforce the reading that Kant viewed himself as inheriting Hume’s wisdom, without needing to defend against any threat Hume might represent.59 The oppositional and inheritor readings are two important traditions in Kant interpre-

55See Stroud(1968/2000), 24 for a general version of this worry, stated with regard to any argument which tries to refute a skeptic by arguing that her skepticism depends on the truth of what she doubts. 56For a version of this criticism, see Guyer(1987), 68. For a rejoinder, on which even a robust account of experience would not beg the question against Hume, see Sommerlatte(2016), 450. 57For a criticism of the Analogies of Experience as “begging the question against Hume,” see Watkins (2004), 485. 58Hatﬁeld(2003), 178. Compare also Ameriks(1978), 273; Watkins(2004) 453, 485–87; Allais(2010), 102. 59For more moderate articulations of the inheritor reading, see Th¨ole(1991), 29; Chance(2011), 327, 342.

109 tation, both with significant textual and philosophical motivation that has kept them alive since Kant’s time. But it would be a distortion to think that, in understanding Kant’s attitude toward Hume, we essentially have two options. For one thing, the two readings are not obviously mutually exclusive. Kant may have wanted to respond to Hume’s skepticism, thereby rescuing certain a priori principles, and with them empirical knowledge, while at the same time refining and more carefully laying the ground for Hume’s censoring of speculative metaphysics. Brian Chance is somewhat unorthodox in suggesting a hybrid reading: “Depending on where one looks, there is. . . ample reason to think that Kant regarded Hume’s philosophy as a cautionary tale and a forerunner to his own. What I have suggested. . . is that we should regard it as both.”60 The conceptual space for this view is easily missed, since each interpretation is often wielded against the other. As we saw, proponents of the oppositional reading have emphasized the difference in goals and philosophical doctrines between the two philosophers, claiming we misunderstand Kant’s intentions, and cannot even understand the content of his arguments, if we do not see him as fundamentally opposed to Hume. Similarly, proponents of the inheritor reading often exploit Kant’s praise of Hume’s censorship of metaphysics as evidence that we should not read his central arguments as in any sense anti-skeptical. I think both criticisms have a ring of truth, if we assume a response to skepticism must not beg the question against the empiricism which underwrites it. But I want to question this assumption, which I think presents a rather narrow view of philosophical dialectic, on which one can succeed at refuting an opponent only with resources that are consistent with and acceptable from the opponent’s initial position. This is not philosophy’s only resource, and it is not the way, I argue, that Kant thought his philosophy could move a skeptic like Hume. That way, I will argue, takes the form of explaining an alternative view of the mind’s relation to the objects of knowledge. Though this explanation may seem to involve claims or perspectives foreign or even antithetical to the skeptic’s view, Kant thought a skeptic may nonetheless be not only able, but motivated to try these out for size. If that is right, then even Kant’s taking for granted that we have experience which already employs the categories can be consistent both with an organizing desire to respond to a skeptical

60Chance(2011), 342.

110 empiricist, and with success in changing a skeptic’s mind. Moreover, we can acknowledge Kant’s negative aims, and his interest in continuing Hume’s dialectical project, while at the same time acknowledging his aim to address Hume’s skepticism and show how to avoid its despairing results. Importantly, then, the oppositional and inheritor readings are also not exhaustive. They are not our only options. Indeed, I think both sides are missing something important here. Proponents of the inheritor reading tend to neglect to ask the question whether Kant thought skeptics would agree with his conclusions in the Analytic. And proponents of both readings have assumed that Kant’s arguments can refute empiricism only if the empiricist accepts their premises while believing that empiricism is true. But that is not how Kant sees things. For Kant sees his Deduction as oﬀering the skeptical empiricist a way out of her skepticism, even if it deploys premises she would not accept while holding fast to her empiricism. This can sound fantastical. But, I will argue, Kant is reasonable to think this, given his understanding of the sources of skeptical empiricism. This makes room for a distinctive third reading, on which Kant’s aim is to change the skeptic’s mind without a non-question-begging refutation. One might, if one is inclined, call this third reading a ‘helpful friend’ reading. How could Kant get the skeptic on board, without a non-question-begging refutation? As a ﬁrst step toward an answer, let us look to Kant’s explicit discussions of skepticism.

4.3 KANT’S ATTITUDE TOWARDS SKEPTICAL EMPIRICISTS

In order to better grasp how Kant hopes to convince a skeptic of his philosophical doctrines, it helps to look closely at what Kant explicitly says about skeptics and about Hume in particular. To those who are inclined to view Kant as the great subverter of skepticism and empiricism, it may be striking to find that his discussions of such figures often have a favorable tone. He sees their skepticism as an important stage along the way to the age of peace in theoretical philosophy that he hoped to usher in with his Critique. Even Kant’s criticisms of skeptics, which for him includes Hume, tend to emphasize what these figures get right, and treat the skeptic’s errors as superable. All of these features of Kant’s portrayals

111 suggest that Kant views skeptics as potential allies, nurturing hopes for mutual cooperation. This suggests, I will argue, that he hoped to correct the skeptics’ errors not by attacking their position as fundamentally flawed, but by explaining an alternative which he thinks they have overlooked. Kant’s first mention of the “sceptic” in the Critique appears within a discussion of the historical development of metaphysics from dogmatism towards “critique”—namely, “a tri- bunal which will assure to reason its lawful claims, and dismiss all groundless pretensions, not by despotic decrees, but in accordance with its own eternal and unalterable laws.”61 Skeptics are here described as reacting against the innumerable and unruly disputes between warring dogmatists. In this regard, the skeptics acknowledge and condemn the lawlessness of the dogmatists’ reigns, just as Kant does. For Kant, their attitude is a sign of increased maturity, albeit one that stagnates in a “doubt” about the possibility of substantive knowledge a priori. Kant initially appears hostile to the skeptics in this discussion. He says, “Happily they were few in number.”62 But elsewhere he sees skeptics as playing an important role in the progress of metaphysics. He describes their attempts at censoring dogmatic metaphysics as warranted and inevitable—“there is really no other available course of action.”63 And he sees the skeptics’ aim as fundamentally the same as his own in the Critique—namely, to “escap[e] from the troublesome affairs of reason. . . [and] arrive at a permanent peace in philosophy.”64 He even credits the skeptic with anticipating, though perhaps only hazily, the critical enterprise through which Kant himself proposes to establish this peace. “The sceptic,” he says, “is thus a taskmaster who constrains the dogmatic reasoner to develop a sound critique of the understanding and reason.” In this way, the sceptic “prepares the way” to critique “by arousing reason to circumspection, and by indicating the radical measures which are adequate to secure it in its legitimate possessions.”65 Kant thus shows a kind of respect for the skeptic, and sees himself as standing on the skeptic’s shoulders. This is especially clear in the case of Hume, whom Kant calls his “sagacious predeces-

61Aix–Axii. See also B23; A760–61/B788–89; P 4:271–75; UE¨ 8:226–28; FM 20:281, 20:342. 62Aix. 63A757/B785. 64A757/B785. 65A769/B797.

112 sor.”66 Repeatedly referring to Hume with the honorable epithet “the acute man,”67 or even “so acute and estimable a man”68 and “the most ingenious of all the sceptics,”69 Kant praises both his intelligence and his character, which Kant takes to be representative of his good judgment. He says that Hume is “well disposed and in his moral character quite blame- less,”70 that his “dispassionate” temperament is “peculiarly fitted for balanced judgment,”71 and, relatedly, that Hume’s writings give “a gentle, calm, unprejudiced examination” of their subjects.72 Kant holds that Hume “struck a spark”73 with his skeptical philosophy and that “something similar to critique of pure reason was found with David Hume.”74 For Kant, Hume was remarkably and uniquely close to giving the survey of reason’s capabilities that Kant intends to give in the Critique. Kant says of Hume, for example, that “deduction. . . had never occurred to anyone but him.”75 And, famously, Kant portrays his own awakening from “dogmatic slumbers,” and turn towards a critical enterprise, as inspired by “the hint that Hume’s doubts had been able to give.”76 Kant repeatedly describes skepticism as a transitional stage on the way to critique. “Scepticism is thus a resting-place for human reason, where it can reflect upon its dogmatic wanderings and make survey of the region in which it finds itself, so that for the future it may be able to choose its path with more certainty. But it is no dwelling-place for permanent settlement.”77 Skeptics like Hume, however, tend to stagnate in a “skeptical stasis.”78 Hume, Kant tells us, was a “geographer of human reason,” who attempted to show that much of our alleged knowledge lies inaccessible “outside the horizon of human reason.”79 But because he did not see how to give a full and exhaustive inquiry into the faculties of reason and

66P 4:260. 67P 4:273, 4:277; KpV 5:53. 68A767/B795. 69A764/B792. 70A746/B774. 71A745/B773. Compare Kant’s estimation of Hume’s character in a 1768 letter to Herder, discussed below on p. 147. 72BL 24:217. 73P 4:257. 74MM 29:782. 75P 4:260. 76P 4:260, 4:262. For discussions of Hume’s inﬂuence on Kant’s theoretical philosophy, see Kuehn(1983); Ertl(2002); Chance(2011); Schafer(forthcoming). 77A761/B789. 78FM 20:281. 79A760/B788.

113 understanding, “Hume. . . foresaw nothing of any such possible formal science [of critique], but deposited his ship on the beach (of skepticism) for safekeeping, where it could lie and rot.” For Kant, however, “it is important. . . to give [the ship] a pilot.”80 This, Kant thought, would require a more thorough inquiry into our faculties than Hume was able to offer. That Hume, and skeptics in general, err, Kant thinks, is clear from the fact that they take their censure of dogmatism too far: “[W]hile rightly denying to the understanding what it cannot really supply, they go on to deny it all power of extending itself a priori.”81 As a result, they call into question various things Kant thinks we know to be true, including the principles we employ in experience, and our mathematical and scientific knowledge.82 Kant’s diagnosis of Hume’s error is that he is “overly hasty” or “premature” in his assessment of our abilities.83 Kant complains that Hume’s negative verdict about the application of pure concepts, and so about the possibility of synthetic a priori knowledge, is made “in spite of his never having tested [the understanding] as a whole.”84 Again, “he did not make a systematic review of all the various kinds of a priori synthesis ascribable to the understanding.”85 Some terminological clarification may be helpful here. For Kant, knowledge is formed when an intuition, which ‘gives’ us an object, is joined with a concept, through which that object is thought. He called this joining of intuition with concept a ‘synthesis,’ insofar as the two elements combine to create something new—namely, our knowledge. The reason a concept is needed is that intuition provides us with an undifferentiated ‘manifold,’ which can only be thought through the unity provided by a concept.86 As the above passage about “systematic review” makes clear, Kant thought there was more than one way for intuitions and concepts to combine through synthesis. What did Hume hastily overlook, so that he neglected to fully test, and so properly discern, the limits of human understanding? What “never occurred to Hume,” Kant says, is the proper “relation of the understanding to experience”—that is, that “the pure concepts of the understanding. . . [are] not [related] in such a way that they are derived from experience,

80P 4:262. 81A767/B795. 82See note7. 83P 4:258. 84A767/B795. 85A767/B795. 86A19–20/B33–34; A50–52/B74–76; A77–79/B103–05; A99; B129–31.

114 but that experience is derived from them.”87 In other words, Hume failed to entertain Kant’s Copernican turn—“the new point of view” which Kant thinks “enables us to explain how there can be knowledge a priori.”88 This revolutionary way of thinking about the relationship between our minds and the objects of our knowledge ﬂips the empiricist’s story, on which objects exist fully formed and independent of the mind, and the mind conforms to such objects when it takes them in through experience. On Kant’s new way of thinking about the mind, it is objects of knowledge which “conform to our knowledge,” “to our faculties” and, in particular, “to our concepts.”89 Kant’s complaint, then, is that the skeptic fails to “test the understanding as a whole” because she fails to consider what knowledge we can have if we view objects as conforming to our faculties. But it is not at all obvious what that latter consideration could amount to. Presumably, Kant would need to say a great deal more about what it could mean for objects to conform to our faculties in order to be met by anything more than the skeptic’s incredulous stare. But if Kant’s diagnosis of the skeptic’s error is on point, then it seems that the appropriate response would be to explain to the skeptic his alternative way of thinking about the mind. Insofar as he can expect to ﬁnd in the skeptic a good-natured comrade, it seems that she would have the patience to listen. If the explanation were to succeed, Kant would thus show her that skeptical empiricism is not the only viable way to think about the mind. If she is given a way of comprehending the overlooked possibility that the understanding’s pure concepts make possible the objects of experience, she could progress beyond her skepticism and to critique, and, in the process, come to understand the knowledge which she previously called into question. That this is how Kant sees things is strongly suggested by the following passage:

We are actually in possession of a priori synthetic modes of knowledge as is shown by the principles of the understanding which anticipate experience. If anyone is quite unable to comprehend the possibility of these principles, he may at ﬁrst be inclined to doubt whether they actually dwell in us a priori but he cannot on this account declare that they are beyond the powers of the understanding, and so represent all the steps which reason takes under their guidance as being null and void. All that he can say is that if we could have insight into their origin and authenticity, we should be able to determine the scope and

87P 4:313; B127. 88Bxix. 89Bxvi–xvii.

115 limits of our reason, but that until we can have such insight any assertions as to the limits of reason are made at random. . . ; but we cannot therefore altogether deny to reason the right to take such forward steps, once we have prepared and secured the way for them by a more thorough preparation of the ground.90

Here, Kant explains that the skeptic’s doubt concerning the principles of the understanding, and the understanding’s power to validly apply concepts more generally, is based in her being “unable to comprehend the possibility of these principles.” But this doubt does not warrant a denial of our knowledge of these principles or of our valid application of concepts, as the skeptic may sometimes be inclined to think. An appropriate response to the doubt, then, would be to provide an explanation of how we can have synthetic a priori knowledge. Presumably, this could remove the doubt. Moreover, the insight such an explanation would provide would, according to the passage above, enable us to determine the true limits of our reason. It would do this in part by explaining the possible “relation of understanding to experience” which “never occurred to Hume.”91

4.4 THE EXPLANATORY READING OF THE DEDUCTION

I will now argue that the primary task of the Deduction is an explanation of just the possibility which the skeptical empiricist overlooks. I call this the explanatory reading of the Deduction. The strongest evidence for the explanatory reading is the language Kant uses in describing the notion of a transcendental deduction in general, and the task of his Transcendental Deduction of the categories in particular. In a section entitled “The Principles of Any Transcendental Deduction,” Kant introduces the notion of a transcendental deduction.92 He says: “the explanation of the manner [Erkl¨arungder Art] in which concepts can thus relate a priori to objects I entitle their transcendental deduction.”93 Transcendental deductions, according to this deﬁnition, essentially “explain the manner” in which concepts relate a

90A762–63/B790–91. 91P 4:313, quoted earlier. 92The title appears at A84/B116. 93A85/B117, my emphasis.

116 priori to the objects of knowledge. An “explanation of the manner” in which certain concepts relate a priori to objects can sound too unambitious a goal to someone who thinks the Deduction is meant to refute certain skeptics. I speculate that this is why the definition of a transcendental deduction has not received due attention. Yet this definition’s portrayal of a deduction as “explanatory” is by no means isolated. Kant regularly described his Transcendental Deduction of the Pure Concepts using similar terms. In the first edition preface, for example, Kant says the De- duction will “expound and render intelligible [dartun und begreiflich machen] [the] objective validity” of the categories.94 The Deduction is there portrayed as explaining the way in which concepts which have their home in the understanding prior to experience relate to experience’s objects. A similar characterization occurs within the first edition’s Deduction. At its beginning, Kant says that “the understanding,. . . as a faculty of knowledge that is meant to relate to objects, calls for clarification [Erläuterung] in regard to the possibility of such relation.”95 At the end of the A-Deduction, Kant says that the goal of “the transcendental deduction of the categories [was] to render comprehensible [begreiflich machen] this relation of understanding. . . to all objects of experience.”96 Such claims also appear in the Critique’s second edition: In a carefully worded, one-sentence “outline” of the second edition’s De- duction, Kant says, “The deduction is the exhibition [Darstellung] of the pure concepts of the understanding, and therewith of all theoretical a priori knowledge, as principles of the possibility of experience.”97 It is worth pausing to ask what kind of explanation Kant thinks a transcendental deduction is. After all, in the quoted passages, Kant uses a diverse set of terms to characterize deductions as explanatory in nature. Some of these expressions are terms of art within

94Axvi–xvii, my emphasis. The ‘und’ here is naturally read as indicating that the two verbs flanking it both characterize the same action, rather than distinct actions. So I agree with Kemp Smith’s rendering of ‘dartun’ in this context as ‘expounding’ rather than ‘proving’ as I think the latter does not characterize the same action as ‘begreiflich machen’ does. Though ‘prove’ may seem to be suggested by the inferential language Kant uses throughout the Deduction, I argue on p. 127 below that such language is consistent with the Deduction’s being primarily an explanation, rather than a proof. 95A97, my emphasis. I read ‘Erläuterung’ as ‘clarification,’ rather than Kemp Smith’s ‘explanation.’ 96A128, my emphasis. 97B168, my emphasis. I read ‘Darstellung’ here as ‘exhibition,’ following Kant’s several explicit identifi- cations between the German and Latin terms (KU 5:192; Anth 7:167). Kemp Smith’s ‘exposition’ is better reserved for Kant’s use of that Latin term or of the German expression ‘Erörterung.’

117 Kant’s critical system, and others have multiple senses. As Kant notes, the German word Erklärung, which figures in Kant’s official characterization of such a deduction, is ambiguous between four different Latin terms: namely, “exposition, explication, declaration, and definition.”98 Kant never says which he has in mind. But two can easily be ruled out. Declaration [Deklaration], for example, seems inapt. For in giving a declaration one “invents” a concept that is “not given to [her] by the nature of understanding or by experience but [instead] is such as [she] has deliberately made it to be.”99 Declaration has a use, for example, in engi- neering, when one wants to create a product with a combination of certain features, but does not yet know if such a product is possible. To borrow Kant’s example, one might declare that the product to be made is a “ship’s clock [Schiffsuhr]” before she knows whether it is possible to make a clock that does not rely on a pendulum, and so does not need the steady ground of dry land to function.100 But insofar as deductions concern the concepts which lie in the understanding a priori, a deduction cannot be a declaration. Nor can it be a definition [Definition]. A definition, Kant says, “present[s] the complete and original concept of a thing within the limits of its concept.”101 But, he continues, we can never be sure that our given empirical and a priori concepts are complete. So definitions turn out to have proper use only in mathematics, whose concepts are fully determined by “an arbitrary synthesis that admits of a priori construction.”102 Philosophy, in contrast, must content itself with “analyses of given concepts,” since it “presuppose[s] the prior presence of concepts, although in a confused state.”103 Kant deems conceptual analysis that is not necessarily exhaustive of its concept ‘exposition’ [Exposition], “a more guarded term [than definition], which the critic can accept” as describing what is actually done in philosophy.104 As the Latin term suggests, such analyses draw or put [-pos-] out [ex-] into the open what is at first hidden away within a concept. Insofar as Kant portrays exposition as playing a role in philosophy and critique, it seems to be a candidate for the kind of explanation which figures in the Deduction. Kant’s identification

98A730/B758. 99A729/B757. 100A729/B757. 101A727/B755. 102A729/B757. 103A730/B758. 104A279/B757. See also A23/B38; JL 9:105.

118 of the Latin term with the German expression ‘Erörterung’ can seem to provide additional evidence.105 This is because it is the term Kant uses in his “transcendental exposition [Erörterung]” of space and time in the Transcendental Aesthetic.106 There, he says his task is to provide an “explanation [Erklärung]” that “makes intelligible” or “renders comprehensible [begreiflich machen]” a body of synthetic a priori knowledge involving spatial and temporal concepts.107 And, tellingly, Kant calls the Transcendental Exposition a “transcendental deduction” of the concept of space.108 The Transcendental Exposition counts as a deduction, Kant seems to suggest, insofar as it “explains [begreiflich machen] how the concepts of space and time. . . must necessarily relate to objects.”109 By the same token, transcendental deductions in general seem to be a kind of exposition—in particular, expositions which draw out the a priori relation of concepts to objects. In the passage where Kant disambiguates the four senses of Erklärung, he says nothing of the remaining sense: explication [Explikation]. But Kant discusses this Latin term in his logic lectures. We see this, for example, when he contrasts empty tautologies (such as ‘man is man’) with “propositions that are identical implicite.” These latter propositions, he says, “make clear the predicate that lay undeveloped (implicite) in the concept of the subject through development [Entwicklung](explicatio).”110 Again, the Latin term suggests the idea of a clarifying or coming to a better grasp of something without actually adding to or extending it—in this case, an unfolding or, more literally, a folding [-plic-] out [ex- ]. The German ‘Entwicklung’ has similar roots—winding [-wick-] away or outward [ent-], i.e., unwinding—and so also suggests a kind of progress or development without actual expansion. Kant does not speak of ‘Explikation’ or ‘Entwicklung’ in connection with the Deduction. But, as noted above, he does describe his Deduction as an ‘Erläuterung’ or ‘clarification.’111 And this term plausibly carries a near identical meaning, insofar as Kant

105A23/B38. The German expression is apt, insofar as the root ‘-¨ort-’, like ‘-pos-,’ connotes the idea of placement. 106B37, B40, B46, B48. 107B41, B48. The knowledge to be explained is geometrical and arithmetical knowledge. 108B119–121. 109B121. On Kant’s portrayal of the Transcendental Exposition as a deduction, see Ameriks(1978), 273– 276. 110JL 9:111. 111A97.

119 sets the related term ‘Läuterung,’ meaning ‘clearing’ or ‘refining,’ over against ‘Erweiterung,’ meaning extending or adding-to. This contrast appears in an important characterization of critique as not itself “a doctrine” of knowledge: “critique[’s]. . . utility, in speculation, ought properly to be only negative, not to extend [erweitern], but only to clarify [läutern] our reason, and keep it free from error.”112 This negative characterization of critique suggests that the kind of explanation given in the Deduction, a crucial step in the critical enterprise, unfolds or unwinds the concepts that reason already has before it a priori. But it does so in a way that extends our knowledge no further than what we already implicitly recognize in the having of such concepts. We do not acquire new a priori knowledge in the process, but instead clarify what we already hazily know, with the effect of guarding us from the error of thinking we may acquire more. The explanation provided in the Deduction, it then seems, is an exposition or explication of what we already recognize about the a priori application of our concepts. Whether it is one or the other, or both, it is to proceed with the aid of “analytic judgments,” since these “really teach us nothing more about the object than what the concept which we have of it already contains; they do not extend [erweitern] our knowledge beyond the concept of the object, but only clarify [erläutern] the concept.”113 But we might wonder how this clarifying is supposed to work. Kant’s aforementioned description of the B-Deduction as an “exhibition [Darstellung]”114 suggests a possible answer: the clarification makes the a priori application of such concepts more ‘tangible’ or ‘intuitable’ to us. The German verb ‘darstellen’ means ‘to exhibit,’ ‘present,’ ‘illustrate,’ or ‘set forth.’ A Darstellung, for Kant, presents its object in a way that makes that object more intuitable or suitable for our sensibility. This connection with sensibility is perhaps more prominent in Kant’s later works, where he describes various parts of our cognitive machinery as perform- ing or being a Darstellung—namely, “the action whereby we give to a concept the corresponding intuition.”115 The power of judgment, for example, performs a Darstellung in the

112A11/B25. Compare Kant’s characterization of critique as the “clarification of our concepts [Aufklärung unserer Begriffe]” (A735/B764). The German for the quote in the body text uses the nominalizations ‘Erweiterung’ and ‘Läuterung.’ I render these in the infinitive above for consistency with Kemp Smith’s translation, which makes the English easier to parse. 113A736/B764. 114B168. 115DWL 24:697.

120 production of knowledge when it “plac[es] a corresponding intuition beside the concept.”116 The power of imagination gives Darstellungen when it constructs determinate concepts out of pure intuition in mathematical knowledge, when it recalls past empirical intuitions for employment in empirical knowledge, and when it presents intuitions without determinate concepts in aesthetic judgment.117 Schemata are “direct. . . presentations [Darstellungen] of the[ir] concepts,” and “make sensible [versinnlichen]” those concepts.118 Kant also uses the term ‘Darstellung’ to represent the way certain art forms make their subjects sensible. So, for example, he says: “Rhetoric can be combined with a painterly presentation [Darstellung] of its subjects as well as objects in a play; poetry with music in song; this, in turn, with a painterly (theatrical) presentation in an opera.”119 In each case, the art resulting from the combination has an additional sense modality through which it can make its subjects more intuitable. Though not yet the term of art it would later become, ‘Darstellung’ already has a relation to comprehension through sensibility in Kant’s ﬁrst Critique. Kant uses the word ‘Darstellung’ and cognates just under a dozen times in the second edition, and often in connection to the enhancement of presentational clarity. Four of the uses occur in the context of Kant’s explaining “changes” he made to the design or “mode of exhibition [Ab¨anderungen der Darstellungsart]” between the two editions of the Critique.120 Kant speaks of Darstellung also in connection with arguments and explanations.121 Discussing a faulty and deceptive cosmological argument for the existence of God, Kant mentions that he will “set out” or “exhibit” the argument in syllogistic form in order to make clear its fallacious and misleading character.122 Elsewhere, Kant speaks of drawing a line as a “mode of depicting [Darstel-

116KU 5:192–93. 117For the ﬁrst, see B156; KU 5:241; Anth 7:167; JL 9:23; DWL 24:697. For the second, see Anth 7:167. For the third, see KU 5:232, 5:244. 118KU 5:351–52. The German uses the nominalization ‘Versinnlichung,’ translating roughly to ‘a making sensible.’ I opt to use the verb form in the body text to avoid the diﬃculty of rendering the noun form in the plural in English. For further discussion of Darstellung, especially in connection to the imagination, see Matherne(2015) and Schubbach(2017). 119KU 5:325. 120E.g., Bxxxviii. 121A608/B636 and A562/B590, respectively. 122“Alle Blendwerke im Schließen entdecken sich am leichtesten, wenn man sie auf schulgerechte Art vor Augen stellt. Hier ist eine solche Darstellung” (A608/B636).

121 lungsart]” the line’s singleness of dimension.123 These are disparate uses, but they seem to share a common element: the idea of arranging something for viewing so it can be seen in a way that makes its structure intuitable. The word ‘Darstellung,’ for Kant, seems to evoke setting forth something tangible for viewing or surveying, as one does in drawing a diagram of a mechanism or laying out a blueprint for the building of a house. Such an exhibition makes the structure of something clear by the arrangement of what it sets forth. Plausibly, an exhibition of this sort is what is requested when Kant asks “how [wie, i.e., in what way],” not whether, “metaphysics as science” or “synthetic a priori knowledge is possible.”124 We might then expect the Deduction to provide something akin to a picture of our cognitive faculties, oﬀered as a way of making intuitable what we already hazily know about their a priori structure and relations. Of course, there is no literal picture or diagram in the Deduction. And the cognitive faculties described by Kant’s transcendental psychology cannot actually be met in pure intuition or experience. So, strictly speaking, we cannot have any intuition corresponding to them. But the Deduction’s descriptions of our faculties is like a picture insofar as it illustrates their a priori functions. Presumably, the notions of synthesis and combination, which Kant borrows from chemistry125 in order to characterize the a priori relations between concepts, intuitions, and objects, play a role; such notions are amenable to being imagined in a sensible way. So, in the end, I think the talk of Darstellung proves apt, especially when we look to the Deduction’s structure, and compare that structure to an explanation aided by a diagram. I show this in the next section.

4.5 THE EXPLANATION IN THE DEDUCTION

In support of the explanatory reading, I have leaned on the words Kant uses to describe the goal or nature of transcendental deductions in general, and in particular the Transcen-

123“. . . als unter dem Bilde einer Linie, sofern wir sie ziehen, ohne welche Darstellungsart wir die Einheit ihrer Abmessung gar nicht erkennen k¨onnten” (B156). 124B19–B22, my emphasis; P 4:276–80. 125For analogies between synthesis in philosophy and in chemistry, see Bxxn and A842/B870. For discussion of these analogies, see Schmid(ms). See also Kant’s praise of chemists such as Ernst Stahl at Bxiii, and talk of acids and alkalis at KpV 5:92.

122 dental Deduction of the categories. But a full defense of this interpretation would need to demonstrate that Kant’s portrayal of the Deduction outside of that stretch of text accurately describes what appears within it. Presumably, this would require going into detail about how the content of the Deduction, or even both editions’ Deductions, is supposed to “exhibit” and “render intelligible” the understanding’s a priori relating to the objects of experience. Here, I do not have the space to advance a detailed, section by section, reading of either of the editions’ Deductions. Instead, my goal in what follows is to focus on what the Deduction so-interpreted can offer to a skeptical empiricist. But before I turn to that, I want to give some further evidence in support of my reading. I do this, first, by observing that reading the Deduction as principally an explanation helps to resolve a puzzle about the structure of the B-edition’s Deduction. Showing how this is so will require wading ankle deep into the content of that edition’s Deduction. But I will only go as far as addressing the puzzle requires. Similarly, I will then go on to briefly consider and reply to an objection concerning the Transcendental Deduction’s relation to the subsequent parts of the Transcendental An- alytic, without pretending to give a complete account of the structure of the Analytic. By addressing certain textual problems, both tasks support my reading, even if leaving further curiosities unaddressed. The textual puzzle concerns why the B-Deduction seems to have two distinct parts which terminate with what, at first pass, can appear to be identical conclusions. In the last sentence of §20 about half of the way through the Deduction, Kant writes: “Consequently, the manifold in a given intuition is necessarily subject to the categories.”126 And, in the last sentence of the substantive portion of §26, and of the Deduction as a whole,127 he says: “All synthesis, therefore, even that which renders perception possible, is subject to the categories.” In the sentence beforehand, Kant identifies the synthesis which renders perception possible as “combination” of the manifold given in “our sensible intuition.”128 And so the conclusion Kant draws in §26 can be restated as follows: The manifold given in our sensible intuition is necessarily subject to the categories.129

126B143. 127That is, before Kant’s discussion of several examples, beginning at B162 and oﬀset by asterisks, and his postmortem reﬂections in §27. 128B161. 129See Henrich(1969), 641–42 for an articulation of this puzzle, which he calls “the problem of the two-steps-

123 It may seem that the puzzle can be resolved easily by acknowledging the addition of the words ‘our sensible’ before ‘intuition.’ This addition indicates that the first conclusion concerns the categories’ application to the manifold given in intuition in general, while the second concerns that application to the manifold given in a certain species of intuition—namely, the species given by human sensibility, whose forms Kant has argued in the Transcendental Aes- thetic are space and time. The claims then differ with respect to their generality. The first is the more general; the second states nearly the same thing, but restricts it to a certain species. But this supposed resolution raises a further question: If the only difference in content between the claims is between their levels of generality, then the first claim implies the second, and can be validly inferred through a simple application of universal instantiation. But, if the second claim follows so obviously from the first, why does Kant not immediately conclude the second at the end of §20? Why does he feel impelled to go on another sixteen pages before reaching the second conclusion in §26? Kant clearly states in §21 that, with the conclusion of §20, “a beginning is made of a deduction of the pure concepts.”130 But why must this conclusion be only a beginning, if it strictly implies what Kant is to conclude much later? The puzzle seems unresolved until this explanation can be given. This puzzle is pressing, if the B-Deduction is conceived as a single, sustained proof or argument. But it can be relaxed and resolved by reading the Deduction as an explanation, rather than an argument. It is natural for explanations of a complicated phenomenon to proceed in stages conducted at various levels of abstraction. Doing so helps to make the complicated explanation more comprehensible. Reading the Deduction as a sustained explanation of single complex phenomenon, proceeding in distinct phases of increasing levels of detail in order to aid in comprehension, then, provides a natural explanation for the Deduction’s multiple parts. Before presenting the textual evidence it will be helpful to consider a more concrete instance of a multi-phased explanation of a complicated phenomenon. Take, for example, an explanation of how a steam locomotive moves along railroad tracks. Such an explanation

in-one-proof.” See also Allison(1983/2004), 160, who calls this “the basic exegetical diﬃculty concerning the B-Deduction: the division of the argument into two parts, each of which presumably establishes the necessity of the categories.” 130B144.

124 might begin by introducing the parts of the steam engine and their connections to each other and to the wheels of the train. For ease of comprehension, this phase of the explanation might abstract away from any details about the generation, channeling, and transfer of steam pressure, focusing instead on how the back-and-forth movement of a piston in the engine is transformed into the rotational movement of a driving wheel. The next phase of the explanation can go on to present the heat and pressure mechanisms responsible for the piston’s motion. This second phase introduces such notions as the burning of coal in the firebox, the heating of water in the boiler, the generation of steam pressure in the main steam pipe, and this pressure’s eventual forcing of the slide valve and piston in the engine’s cylinder. The first phase of the explanation, which abstracts away from such steam pressure mechanisms, provides the spatial and mechanical context for understanding their role in moving the locomotive forward. Though both phases give an explanation of how the locomotive moves forward, the first phase of the explanation is not yet unique to a steam locomotive. It is more generic, and could explain the movement of any wheeled vehicle with a reciprocating piston engine. So the explanation remains incomplete until the second phase’s fuller account. But presenting both phases at once may overwhelm, and the two are naturally separable based on the different principles of mechanics most central to each: kinetics in the first; changes of matter and gas-pressure laws in the second. I believe we can find a similar structure of explanation in the Deduction. Its overall goal is to explain or render intelligible the categories’ application to the objects of experience. Since those objects are provided by the mind’s receptivity—and, in particular, by empirical intuition—the explanation turns out to be a story about how applying the categories to the material provided by empirical intuition makes that material thinkable. Kant calls this act, the act of making intuited material thinkable, the combination of the manifold in a given intuition. That is why the notion of combination figures in the conclusions he draws in §§20 and 26. If we can conceive of the application of the categories as constituting the act of combining the manifold, and conceive of the intuition which provides the manifold as human empirical intuition, then the deduction provides a sketch of how the categories apply to the objects of our experience, and indeed make this experience possible in the first place. The application of categories is a very complex phenomenon, and not easily grasped

125 all at once. So Kant introduces it in several phases, each proceeding at different levels of abstraction, though all focusing on the central notion of the combination of the manifold of a given intuition. I say ‘several’ phases, because I think the Deduction is naturally read as involving three such phases, occurring primarily in §§16–20, 24, and 26, respectively. The remaining sections of the Deduction step back from the body of the explanation to comment on its structure, clear up tangential issues, or, most important to Kant’s aims in the Critique, draw conclusions concerning the possibility and proper objects of metaphysics.131 The first and most involved phase explains the act of combining the manifold in abstraction from the particular forms or modes of human sensibility. After introducing the notion of combination as the act responsible for unifying the manifold in §15, Kant in §16 explicates the notion of an ‘original’ combination in virtue of which one can become conscious of all one’s representations in thought as one’s own. In §17, Kant identifies this combination with the act responsible for one’s thoughts’ relating to objects. He then clarifies that what allows this act of combination to relate to objects is the application of a general concept—that is, that the manifold of a given intuition is unified in the concept of an object. In §18, Kant contrasts the unity brought about by the application of the concept of an object with a subjective unity of consciousness, in order to bring out the sense in which the original combination is objective and prior. In §19, he explains that the very same act of combination, now conceived in terms of its unification of the manifold in the concept of an object, just is the logical function of judgment. Since, as Kant found in the preceding Metaphysical Deduction, these logical functions just are the categories insofar as the former are employed in the determination of the manifold given in intuition, he concludes in §20 that the manifold is necessarily subject to the categories. In these sections, a single act of combination is held up to many lights—roughly, it is viewed in terms of its role in our consciousness and thought (§16), in knowledge (§§17–18), in judgment (§19), and with regard to intuition (throughout, though especially §20). Each new perspective fills in a part of the picture. But throughout this whole discussion thus far, Kant observes in §21, the way in which the manifold is intuited “remains here undetermined.”132 Kant explains that he “must abstract from the mode in

131See Edgar(2010) for a reading along these lines, motivated as a response to the same puzzle. 132B145.

126 which the manifold for an empirical intuition is given in order to direct attention solely to the unity which, in terms of the category, and by means of the understanding, enters into the intuition.”133 Doing so, he suggests, helps “direct attention” toward the act of combination responsible for the unity of the intuition, and how the categories figure in this act. Issues about how human empirical intuition figures in the picture are prudently postponed. Indeed, this phase of the explanation, for which my quick sketch draws only broad contours, is demanding enough to comprehend as it is. At this point, the explanation applies equally to any finite knower, whose knowledge involves the application of concepts to given intuitions.134 But just as we may want to understand the workings of a steam locomotive, as opposed to vehicles with engines and wheels more generally, we may want to know how human empirical knowledge, as opposed to the knowledge of finite knowers in general, presupposes application of pure concepts. A further explanation would show how such concepts apply to the objects given through human sensibility—that is, objects in space and time. And only this further explanation would show that our application of the categories makes experience possible. But this part of the explanation, Kant explains in §22, itself involves two phases, insofar as “sensible intuition is either pure intuition (space and time) or empirical intuition of that which is immediately represented, through sensation, as actual in space and time.”135 The first of these latter two phases, which occurs in §24, explains how pure intuition is combined by the application of the categories. This phase begins to explain the possibility of mathematical knowledge. But it does not complete this explanation insofar as the combination of pure intuitions “yields knowledge only in so far as these intuitions. . . can be applied to empirical intuition.”136 The final phase, which occurs in §26, then completes the explanation by explaining how the application of the categories combines the manifold given in empirical sensible intuition. This reading may seem hard to square with the inferential language Kant uses throughout the Deduction. We have already seen such language in the articulation of our puzzle. Kant

133B144. Here, reading ‘um. . . zu’ as ‘in order to’ rather than Kemp Smith’s ‘and must’. 134The explanation “could not abstract [from the fact]. . . that the manifold to be intuited must be given prior to the synthesis of understanding and independently,” for to do so would change the subject of explanation to that of “a divine understanding” whose representations are never aﬀected by objects but instead produce them” (B145). 135B147. 136B147.

127 seems to portray himself as drawing conclusions at §§20 and 26. But here we may observe that saying that the Deduction is an explanation, rather than an argument, need not deny that the explanation contains arguments within it. It is normal for explanations to do this. If I have explained that the amount of the steam pressure which travels to the pistons can be manually controlled by a regulator valve, I might then conclude by saying that the speed at which the locomotive travels, thus, can be controlled by adjusting the angle of the regulator valve. The inferential language here would then indicate that understanding why the claim is true depends on recalling connections already made earlier in the explanation— in this case, the steam pressure’s sliding a piston, whose movement is then transferred to the rotations of the wheels. Kant’s conclusions in the Deduction are like this. They prompt the reader to recall sequentially introduced connections between the diverse parts of a unitary phenomenon. In this way, it is not strange for Kant to say that the Deduction “explain[s] how there can be knowledge a priori” and is one of Kant’s “proofs of the laws which form the a priori basis of nature” in nearly the same breath.137 Again, what I have provided so far are just the basic outlines of a reading of the Deduction as a multiphase explanation. A more complete defense would require close readings of each phase, and would more properly be the subject of a book-length project. For our purposes it will suffice to say that the reading finds a suggested confirmation in the first paragraph of §27, which provides a summary of the Deduction. Kant says:

We cannot think an object save through the categories; we cannot know an object so thought save through intuitions corresponding to these concepts. Now all our intuitions are sensible; and this knowledge, in so far as its object is given, is empirical. But empirical knowledge is experience. Consequently, there can be no a priori knowledge, except of objects of possible experience.138

In the first sentence, as in the first phase of the explanation, Kant focuses on the application of the categories to intuition in general. In the next clause, he notes that intuition is sensible, specifying the forms of the intuition to be space and time; and, in the next, Kant notes that the intuition must be empirical, insofar as it provides us with objects. These specifications accord with the second and third phases of the Deduction’s explanation as I have spelled

137Bxix. 138B166, Kant’s emphasis.

128 it out above. Noting the equivalence of empirical knowledge with experience, we can then see all three phases as coming together to explain what Kant previously identiﬁed as the “principle according to which the [Deduction’s] whole enquiry must be directed, namely, that [the categories] must be recognised as a priori conditions of the possibility of experience.”139 From the explanation given, Kant draws a negative conclusion about the limits of our a priori knowledge: It can only concern the objects of possible experience. But a positive conclusion follows as well: We actually have a priori knowledge, by applying the categories to the objects of possible experience. Hence, Kant sums up the Deduction at the end of §27 by saying:

The deduction is the exposition [Darstellung] of the pure concepts of the understanding, and therewith of all theoretical a priori knowledge, as principles of the possibility of experience.140

This way of construing the Deduction’s task can lead us to wonder why Kant thinks he has completed it by or with §27. Does what he says in the Deduction truly explain the theoretical a priori knowledge that is to be the subject of the next part of the Transcendental Analytic, the Analytic of Principles? Even Kant can seem to admit that it does not. For he seems to portray the Analytic of Principles as continuing what, according to my interpretation, the deduction is meant to do. He says, “How [the categories] make experience possible. . . will be shown more fully in the following chapter on the transcendental employment of the faculty of judgment.”141 If the job of the deduction is to explain how the categories make experience possible, why does the deduction end before the “fuller” explanation that Kant promises in the Analytic of Principles is given? This challenge is addressed by noting that the deduction does not concern the categories individually, but rather as an interconnected system of so far undiﬀerentiated concepts. When Kant introduces the notion of transcendental deduction in §13, he does so at a high level of generality. The title of the section is “The Principle of Any Transcendental De- duction.”142 At this level of generality, Kant has not yet even distinguished the a priori

139A94/B126. 140B168. 141B167. 142A84/B116, my emphasis.

129 concepts of space and time from the categories, as evident from his talk of a transcendental deduction of the concept of space.143 Once Kant begins §14, he has narrowed his topic to a “Transcendental Deduction of the Categories.”144 But Kant does not specify the categories into the individual twelve listed on his table,145 with the exception of two brief mentions of ‘cause,’146 both of which are clearly used to illustrate a general point applying equally to every category. In short, the Deduction explains how the categories, qua system of so far undifferentiated pure concepts, apply to the objects of knowledge. Given the level of abstraction at which the deduction operates, it is natural to expect that more explanation may be desired. One can wonder, for example, how it is that the individual categories—substance, cause, and community, for example—can “subsume intuitions,”147 which, according to Kant, they must in order to yield experience of objects in space and time. Explaining how this occurs is the task of the schematism, which designates a schema corresponding to “each category.”148 Here, unlike in the Deduction, the categories are treated individually, rather than as undifferentiated elements in an interconnected system. A similar point can be made with respect to the pure principles of the understanding. The analogies of experience, for example, concern the individual categories of ‘substance,’ ‘cause,’ and ‘community.’149 This focus on the individual categories is, however, not the only difference in aim between the Deduction and what comes after. Kant registers another difference at the beginning of the Analytic of Principles when he mentions that “the preceding chapter” concerned the “universal conditions under which [judgment] is alone justified in employing pure concepts of understanding,” while “our task now is to exhibit. . . the judgments which understanding, under this critical supervision, actually achieves a priori.”150 The Deduction does not seek to address what judgments the understanding can form a priori, but rather to explain its possession of the system of concepts which make possible their employment in such judg-

143A88/B120–21. 144A92/B124. 145The table appears at A80/B106. 146B163 and B168. 147A138/B177. 148A143–46/B183–85. 149See A182/B224, A189/B232, and A211/B256, respectively. 150B187.

130 ments. But the Analytic of Principles still continues the task started in the Deduction of explaining how the understanding makes experience possible. At least some of these judgments continue to shed light on what experience must be like in terms of the application of the categories. And their actual employment in experience makes empirical knowledge possible.151 In summary, the Deduction is an explanation, and in particular, either an exposition or an explication. In either case, it clariﬁes what we already, if only hazily, recognize in our concepts. It explains the a priori application of concepts to the objects of experience through a Darstellung, or a making intuitable of this relation. This reading does justice to Kant’s remarks about the nature of a deduction. It helps to resolve exegetical puzzles about the structure of the Deduction and its relations to the later chapters of the Transcendental Analytic. And if I am right, the explanation it gives is an explanation of a possible relation between our minds and the objects of knowledge that is overlooked by the skeptical empiricist.

4.6 THE INSTABILITY OF SKEPTICAL EMPIRICISM

The Deduction of the Pure Concepts oﬀers the skeptical empiricist a way out of his skepticism by providing him with an explanation of a possible relationship between the mind and the objects of knowledge which he overlooks. If he listens to this explanation, he could come to see how pure concepts can validly apply to such objects. In doing so, he would come to see how concepts which originate in the understanding prior to experience can ﬁgure in substantive knowledge about the objects of experience. By coming to see that and how substantive a priori knowledge is possible, he would also come to see how we can have the ordinary empirical knowledge he called into question. He would thus cease to be an empiricist and cease to be a skeptic. If the skeptical empiricist is receptive to Kant’s explanation, it can cure his skepticism. But it can seem hopelessly na¨ıve to think that he will be. The Deduction explains how

151See B218.

131 concepts apply a priori to the objects of experience. But an empiricist proposes that we view all of our thought as arising from the materials of experience. Wielding this methodological commitment, he can reject all alleged substantive a priori knowledge as nonsense or pretense, and claim that we can know nothing of the sort. Moreover, the commitment would require rejecting the Copernican turn on which Kant’s explanation in the Deduction relies. As long as the empiricist holds fast to his empiricist methodology, and embraces the skepticism to which it gives rise, it seems he will find nothing appealing in Kant’s explanation. An explanation of the valid application of concepts not derived from experience must strike him as backwards and vain, if even intelligible. Worse yet, insofar as Kant offers his explanation as a means to progress beyond skepticism, Kant may appear to the empiricist as a sort of snake oil salesman, peddling supposed cures for a condition the empiricist does not yet recognize to be a disease. As long as Kant appears in this guise, his offer of a cure for skepticism will fall on deaf ears. If that is right, the interpretation that Kant tries to change the skeptical empiricist’s mind by offering such an explanation can seem to cast Kant in a poor light. Surely a thinker as shrewd as Kant would have recognized that the explanation given in his Deduction is diametrically opposed to the skeptic’s empiricism. If he saw this, should he not have also realized that he was begging the question against the skeptical empiricist? Expecting her to abandon her empiricism in order to embrace Kant’s alternative may seem tantamount to expecting a sudden and inexplicable conversion. If this was Kant’s expectation, it may then seem that he utterly failed to grasp who the skeptical empiricist is, and why she holds the commitments she does. I want to argue that the opposite is true. Kant’s hope that he could change a skeptical empiricist’s mind reflects a deep and subtle understanding of skepticism and its sources. As we shall see, Kant’s most sustained and penetrating treatment of skeptical empiricism appears in a rich yet underexplored passage within the Critique’s Discipline of Pure Rea- son, entitled “The Impossibility of a Sceptical Satisfaction of Pure Reason in its Internal Conflict.”152 Here, Kant offers a rationale for thinking even the most determined skeptical empiricist must be partly dissatisfied with her position in a way that makes Kant’s explana-

152A758–69/B786–97.

132 tion necessarily appealing to her. Kant reinforces this picture of the skeptic throughout his theoretical works and lectures. For Kant, the skeptic must be dissatisfied with his skepticism, qua reasoner. According to Kant, the drive to ask and answer metaphysical questions is a basic and inevitable “natural disposition” of human reason.153 We humans are “impetuously driven by an inward need to questions such as cannot be answered by empirical employment of reason.”154 Again, “human reason. . . is burdened by questions which, as prescribed by the very nature of reason itself, it is not able to ignore.”155 This “inward need” or “drive” or “itch”156 of reason encourages us to answer substantive questions that cannot be resolved by appeal to any experiences we have had, or even could have. It especially urges us to form judgments concerning the object of traditional or ‘transcendent’ metaphysics—namely, God, the immortal soul, and the freedom of the will—knowledge of which is the “ultimate aim to which the speculation of reason in its transcendental employment is directed.”157 Again, it might seem that the empiricist has grounds for quieting these urges. Hume reports that “despair has almost the same effect upon us with enjoyment, and that we are no sooner acquainted with the impossibility of satisfying any desire, than the desire itself vanishes.”158 If that is right, finding out that it is impossible to answer questions whose resolution demands insight beyond what experience can deliver would destroy our desire to answer such questions. Insofar as Hume’s empiricist principles trace all our ideas back to experience, those principles can seem to prove the impossibility of satisfying such a desire. This suggests that the drive to metaphysics should vanish along with the adoption of skeptical empiricism, or at least once one has worked out the implications of that position. The skeptic could then set aside questions, for example, about God’s existence or life after death as meaningless or unanswerable, and either way best ignored.

153B21; B22. For Kant, we share this dispositions with even primitive humans: “Thus in all men, as soon as their reason has become ripe for speculation, there has always existed and will always continue to exist some kind of metaphysics” (B21). That perhaps explains why metaphysics is, according to Kant, “older than all the other sciences” (Bxiv). “The idea of such a science is as old as speculative human reason” (A842/B860). 154B21. 155Axii. 156TeFP, 8:414. 157A798/B826; cf. A3/B7, B395n. On the drives or needs of reason, see Willaschek(2018), 18, 23–25. 158T Intro.9, SBN xviii.

133 One would be hard pressed to find someone who agrees with Hume about the ease with which our desires vanish. We do not stop wanting to talk to our parents as soon as they die, or stop wanting chocolate after eating the last piece. As we saw in Chapter3, Hume’s own “despair” in the Treatise does not succeed at removing his desire to continue his inquiries. Similarly, for Kant, skeptical empiricism can never succeed at purging our desire to answer metaphysical questions. He says that “to allow ourselves simply to acquiesce in [skeptical] doubts. . . is a futile procedure, and can never suffice to overcome the restlessness of reason.”159 If that is right, the empiricist who claims to have overcome the desire to seek out metaphysical knowledge on the grounds of his empiricism is putting on a pretense. His professed lack of interest stands in tension with the necessary interests of reason within him. Whether or not he sees this with clear eyes, “it is idle to feign indifference to such enquiries, the object of which can never be indifferent to our human reason.”160 “Nimble reason” may break free from the skeptic’s futile attempts to constrain it, and rekindle an interest in metaphysics which he has never fully abandoned.161 Kant’s rationale for thinking that the skeptical empiricist cannot succeed in purging the desire to answer metaphysical questions is not simply that this urge is ineradicable. It is instead that the skeptical empiricist has insufficient grounds for thinking that metaphysics is impossible. To see why Kant thinks this, it is helpful to note that he found skeptical empiricism, and in particular the “sceptical teachings” of Hume, to be “based on facts which are contingent, not on principles which can constrain to a necessary renunciation of all right to dogmatic assertions.”162 These “facts [Fakta],” Kant tells us, concern “all unsuccessful dogmatic attempts of reason.”163 Kant was not often a careful reader of other philosophers, and it is doubtful that he was able to read much of Hume’s Treatise.164 But I believe he here characterizes the grounds of

159A757/B785. 160Ax. 161TeFP 8:415. 162A767–68/B795–96. 163A763–64/B791–92. 164Notes from Kant’s logic lectures from the early 1770s mention and recommend Johann Georg Sulzer’s 1750s, four volume translation of Hume’s works, including the An Enquiry Concerning Human Understand- ing, to those who “want closer instruction in the scepticismus of modern times” (BL 20:217–18). But Sulzer’s translations did not include the Treatise. Kant, whose English was poor, likely had relatively little knowledge of the Treatise. Notable exceptions were Hamann’s 1771 publication of T 1.4.7 under the title “Nachdenken

134 Hume’s philosophical approach quite similarly to Hume’s own portrayal of them in the introduction to the Treatise. There, Hume describes a sorry state of “metaphysical reasonings,” a “bustle” full of “noise and clamour” in which “disputes are multiplied.”165 These embar- rassments in philosophy, for Hume, motivate the need to try out a new method. Noting the explosive successes in the natural sciences in the century following Bacon’s experimentalism, Hume proposes to follow a line of recent English empiricist philosophers in “the application of experimental philosophy” to philosophical topics.166 If philosophy is going to progress beyond needless bickering and talking past one another, Hume suggests, the “science must be laid on experience and observation.”167 In short, Hume seems to adopt an experimental method on the grounds of his observations of the failures and obscurities of past metaphysicians and of the successes and precision of the natural sciences.168 If that is right, Kant would be correct to view Hume’s method, and the “sceptical teachings” which its application leads to, as based upon observations about the failures of dogmatic metaphysics. In doing so, Kant would perhaps be underemphasizing the extent to which observations about the successes in the sciences were among the “facts” which motivated Hume’s adoption of his method. But Kant would be right nonetheless to see the empiricist’s skepticism of metaphysics to be “based only on facts which are contingent,” insofar as that skepticism results from the application of a method adopted in response to facts about how the history of philosophy and the sciences has played out so far. For an empiricist like Hume, who has judged that dogmatic metaphysics ends in obscurity, while the experimental method leads to advancement in the natural sciences and to clarity

eines Zweiﬂers,” and forty scattered quotations, many from T 1.3, in the 1772 German translation of James Beattie’s Essay on the Nature and Immutability of the Truth. More speculatively, Kant may have gained some degree of insight into other parts of the Treatise through conversation with Hamann. For further discussion of the early reception of Hume’s texts in the German-speaking world, see Gawlick and Kreimendahl (1987). 165T Intro.2–3, SBN xiii–xiv. 166T Intro.7 and 7n, SBN xvii. 167T Intro.7, SBN xvii. 168If Kant never discovered the grounds for Hume’s method in the Treatise, he could still have recognized those grounds in the ﬁrst Enquiry. There, too, Hume registers his disappointment with the volumes of “di- vinity and school metaphysics” which contain neither mathematics nor “experimental reasoning concerning matter of fact and existence,” and, as a result, “contain nothing but sophistry and illusion” (EHU 12.34, SBN 165). With the adoption of an experimental method, Hume “hopes, that [through his philosophy] men may reach a greater clearness and precision in philosophical reasonings, than what they have hitherto been able to attain” (EHU 7.4, SBN 62).

135 in philosophy, the experimental method itself rests on good experimental grounds. So if some experimental reasoning were to lead an empiricist to conclude that we cannot gain metaphysical knowledge, he would have what he would consider to be sound evidence for that conclusion. But it is hard to see what sort of experimental reasoning could prove that we cannot ever gain such knowledge. Hume writes:

’Tis impossible to reason justly, without understanding perfectly the idea concerning which we reason; and ’tis impossible perfectly to understand any idea, without tracing it up to its origin, and examining that primary impression, from which it arises.169

The failure of past attempts may very well be evidence that future attempts too will fail. But it does not show that they must. The skeptical empiricist is not entitled to say, as Hume does here, that success in reasoning is “impossible” without a basis in experience. We might still wonder whether another form of reasoning can succeed, despite our failures so far. For Kant, any conclusion based on empirical grounds leaves room for this kind of doubt to creep in: “Thus the fate that waits upon all scepticism likewise befalls Hume, namely, that his own sceptical teaching comes to be doubted, as being based only on facts which are contingent, not on principles which can constrain to a necessary renunciation of all right to dogmatic assertions.”170 The empiricist’s commitment to the experimental method is shaky, Kant thinks, being itself ﬁrst formed in response to contingent, empirical evidence. This instability is inherited by the skepticism to which that method leads. Moreover, Kant thinks we could never know the impossibility of substantive knowledge which is independent of experience. He says:

Nothing worse could happen to [the development of systems of synthetic a priori knowledge] than that someone should make the unexpected discovery that there is and can be no [synthetic] a priori knowledge at all. But there is no danger of this. It would be tantamount to someone’s wanting to prove by reason that there is no reason. For, we say that we know something by reason only when we are aware that we could have known it even if it had not presented itself to us as it did in experience. It is an outright contradiction to want to extract necessity from an empirical proposition (ex pumice aquam) and to give a judgment, along with necessity, true universality.171

169T 1.3.2.4, SBN 74–75. By “impressions” Hume means our “sensations, passions and emotions,” all elements of our experience (T 1.1.1.1, SBN 1). 170A767–68/B795–96. 171KpV 5:12.

136 For Kant, the conclusion that synthetic a priori knowledge is impossible would have to hold a priori, given its modality and universal scope: “experience never confers on its judgments true or strict. . . universality.”172 The conclusion is also synthetic, since the concept of synthetic a priori knowledge does not already contain the concept of impossibility. So the conclusion would be synthetic a priori, and so provide a counterexample to itself, if known. It thus cannot be known. We should not mistake this argument to be a sort of refutation of skeptical empiricism, put forward by Kant to compel the skeptical empiricist to recognize the inconsistency of her own position. I think that Kant would have understood the skeptical empiricist as agreeing with him that we do not know the impossibility of knowledge which is independent of experience, despite her tendency to occasionally assert this impossibility.173 Instead, he sees the empiricist as concluding that metaphysics has poor prospects on the probable evidence of past experience. Kant’s argument does not challenge this claim.174 Instead, the argument reinforces Kant’s point that the empiricist’s skepticism about metaphysics is “based only on facts which are contingent,” and so is liable “to be doubted.”175 The fact that attempts at metaphysical systems have so far failed does not conclusively oblige us to lose hope. For Kant, no empirical finding could. Similarly, experience can never prove that we can never form substantive knowledge independently of experience. Reflection on past experience may suggest that it is improbable that we will form such knowledge. But experience cannot deliver knowledge of its impossibility. Nor can can a priori reasoning, as Kant has just argued. So as long as the skeptical empiricist doubts the possibility of synthetic a priori knowledge, she effectively doubts that this possibility can be conclusively ruled out. Since empirical grounds are insufficient for concluding that we cannot know, Kant thinks, the right response at this point is to continue inquiring. As Kant puts it: “The consciousness of my ignorance (unless at the same time this ignorance is recognised as being necessary), in-

172B3–4; A734/B762. 173Thus, if he had read it, Kant may well have seen Hume’s “’Tis impossible,” quoted above, as an over- statement of his position. 174In fact, Kant’s argument here does not seem to be addressed to a skeptical empiricist at all, since he says just a couple pages later that “in this philosophical and critical age empiricism can scarcely be taken seriously” (KpV, 5:14). Presumably, the ﬁrst Critique helped usher in this age. 175A767–68/B795–96.

137 stead of ending my enquiries, ought rather to be itself the reason for entering upon them.”176 Moreover, as we saw, for Kant, reason itself poses metaphysical questions. On his view, this implies that only reason can quell its metaphysical drive, either by satisfying it through the resolution of such problems or else by discovering its own incapacity to resolve them. It follows that “we have no right to ignore these problems. . . and. . . on the [skeptical empiricist’s] plea of our incapacity, decline to occupy ourselves with their further investigation, for since reason is the sole begetter of these ideas, it is under obligation to give an account of their validity or of their illusory, dialectical nature.”177 Consciousness of our own ignorance about matters in which, as reasoners, we are inherently interested, rightly drives us to continue inquiring into them. Recognizing this, “reason insists on giving free rein to itself,” having “not in the least been disturbed, but only temporarily impeded” by the skeptical empiricists’ assaults.178 This natural drive helps explain why Kant’s explanation of our a priori knowledge must appeal to the skeptic. The skeptical empiricist, as Kant sees her, is necessarily at war with herself. As a skeptical empiricist, she has given up on reason’s inquiries into the possibility of what Kant calls synthetic a priori knowledge, albeit in an empirically grounded way which leaves room for a hope of understanding the possibility and possession of such knowledge. But as a reasoner, she remains interested in whether and how she might possess synthetic a priori knowledge. The German word ‘zwei,’ or ‘two,’ at the root of ‘Zweifel,’ or ‘doubt,’ provides a helpful piece of etymology here. On Kant’s view, the skeptical empiricist is essentially of two minds. She is doubtful and frustrated with respect to the possibility of synthetic a priori knowledge, yet still harboring a perhaps submerged hope and curiosity. Exploiting a further etymological connection, Kant thus characterizes the skeptical empiricist’s doubt [Zweifel] as a kind of “despair [Verzweiﬂung] as regards satisfaction of reason’s most important aims.”179 Elsewhere, he says that “despair [Verzweiﬂung] is a temporary dementedness in someone who is hopeless.”180 This characterization suggests a proper

176A758/B786. 177A763/B791. 178A768/B796. 179P 4:271. Compare “skeptical hopelessness [skeptischen Hoﬀnungslosigkeit]” at A704/B434. Kemp Smith translates this as “skeptical despair,” and reasonably so given Kant’s connection between despair and hopelessness. 180VKK 2:268–69.

138 cure for the despairing, and so too for the skeptic: If such a person can be inspired once again to hope, her “temporary” condition can be left behind. As Kant sees it, empiricists, such as Hume, are struck by a repeated failure to successfully answer metaphysical questions that arise from the nature of reason itself—so struck that, on the basis of these observed failures, they “despair” about the possibility of success, without a proof that success is impossible. It follows that, in some sense, the skeptical empiricist is always ready to give up her skepticism, if the possibility of our possessing such knowledge were to become clear. That could restore the hope which she has lost. What Kant oﬀers in his Deduction, then, promises to give her what she really yearns for—an illustration of the possibility of the natural enterprise that she prematurely abandoned only out of frustration. Nothing, then, locks the skeptical empiricist into her view beyond her despairing doubt. Although previous failures lend some initial support to that doubt, that support is inconclu- sive, and in conﬂict with her natural “drives” as a reasoner. In these two ways, her skepticism is unstable. She may then be drawn out of her despair by an explanation that satisfyingly “renders intelligible” synthetic a priori knowledge. Kant’s Deduction presents an illustration of a possible relationship between concepts and objects which he thinks empiricists like Hume have so far failed to entertain. It is reasonable to think, then, that Kant’s illustration of this relationship could engage the skeptic’s imagination, and get her to comprehend the possibility that she previously did not see. This could allow her to progress beyond her skepticism. In the end, Kant does not think that we can gain knowledge of the traditional topics of metaphysics. According to him, showing this is the primary “utility” of his critique, which “ought properly to be only negative, not to extend, but only to clarify our reason, and keep it free from errors.”181 Insofar as Kant’s Deduction is the most important “enquiry” for “determining the rules and limits of [the] employment” of our reason and understanding, it is to play a central role in explaining where our knowledge must come to an end.182 Indeed, it is at the close of the Deduction that Kant concludes that “there can be no a priori knowledge, except of objects of possible experience.”183 If we can know of nothing beyond the bounds of

181A11/B25. 182Axvi. 183B166.

139 all possible experience, we cannot have knowledge of the traditional objects of transcendent metaphysics: God, freedom, and the soul. So, in the end, Kant’s Deduction cannot satisfy any hope we might have to gain metaphysical knowledge of such objects. But unlike skepticism, the Deduction can, Kant thinks, succeed at purging the desire for such metaphysical knowledge. It does this “by sufficiently clarifying our concepts to recall [reason] from its presumptuous speculative pursuits to modest but thorough self- knowledge.”184 This clarification is the core of what Kant calls critique, a task reason itself undertakes in pursuit of “self-knowledge” about the limits of its own rightful use.185 Through critique, reason comes to know what it can and cannot know a priori. In determining what reason must necessarily remain ignorant about, then, “sober critique” provides “a true cathartic” for our desire to form beliefs about things beyond the limits of all possible experience, thus “effectively guard[ing] us against such groundless beliefs.”186 Reason can then finally dismiss its desire to answer certain metaphysical questions on the basis not of experience of past failures but of knowledge of its own operations. This seems to be Kant’s point when he says that his critique of metaphysics, unlike skepticism’s censure, is “based on principles.”187

184A735/B763. 185Axi. Critique is to produce “scientiﬁc and completely convincing self-knowledge” (A849/B877). 186A486/B514. 187A761/B789; P 4:260–61, 4:270. According to Kant, even critical self-knowledge cannot prevent “transcendental illusion,” which “exerts its inﬂuence on the principles” and “carries us altogether beyond the empirical employment of categories” (A295/B352). In other words, even after determining the limits of our knowledge, we remain tempted to apply pure concepts beyond those limits to transcendent things like God or the world taken as a whole. In this regard, even after critique it can seem to us, for example, that the principle that all things that happen have a cause is applicable to transcendent things, and so that the world taken as whole must have a cause. It continues to seem to us, then, that certain answers to metaphysical questions are correct. “Transcendental illusion. . . does not cease even after it has been detected and its invalidity clearly revealed by transcendental criticism” (A297/B353). The existence of transcendental illusion may seem to show that Kant’s “true cathartic” has failed, insofar as we remain tempted to answer metaphysical questions even after critique. But at least two replies are available to Kant. Kant could say that the temptation of transcendental illusion is merely that of any ‘seeming,’ and need not imply the existence of a desire. The temptation relevant to the illusion of a stick’s being bent in water need not imply that we desire to determine whether the stick is bent. Alternatively, even if transcendental illusion implies the return of a desire to answer metaphysical questions, the cathartic of criticism could always be applied to remove the desire again. Kant thinks that, when we are faced with illusion, knowledge of its illusory nature can keep us from giving in to it and making judgments on its basis. This is analogous to how “the astronomer can[not] prevent the moon from appearing larger at its rising [when it is near the horizon], although he is not deceived by the illusion” (A297/B354). Critique allows us to recognize the illusion for what it is, and so can deprive it of its pull. That could allow us to put to rest, for a while anyways, any desire to answer questions of transcendent metaphysics on its basis. For Kant, even if criticism does not inoculate us against such desire, it is a better medicine than the skeptic’s prescription

140 Still, Kant’s explanation of our knowledge in the Deduction does at the same time give the skeptic a way to comprehend our possession of some synthetic a priori knowledge of objects— namely, the principles of the understanding which we employ in all empirical experience.188 Though Kant goes on to enumerate these principles only after the Deduction, the Deduction helps to explain how we can know them. It does this by explaining the valid application of the pure concepts employed by these principles to the objects of experience. These a priori principles include, for example, the principle that “all alterations take place in conformity with the law of the connection of cause and eﬀect”189—a principle which Kant sees the skeptical empiricist as casting doubt upon. As a result, the Deduction explains the possibility of our knowing what the skeptic doubts. The skeptic’s acceptance of this explanation can remove her doubt, and so lead her out of skepticism. I think the fact that the Deduction explains the possibility of such knowledge can be an additional way that it can appeal to a skeptical empiricist. As Kant sees it, the principles we employ in experience are recognized to be true in our common employment of them. This is suggested, for example, in the Third Analogy, where Kant discusses the principle of coexistence in accordance with reciprocal laws. He says:

We may easily recognise from our experience that only the continuous inﬂuence in all parts of space can lead our senses from one object to another. . . We cannot empirically change our position, and perceive the change, unless matter in all parts of space makes perception of our position possible to us.190

If we indeed “easily recognize” the principles of the understanding by employing them in experience, it seems that even common reason has knowledge of them and can become conscious of them without the aid of philosophy. As Kant puts it: “An a priori proposition

of doubt. The latter not only has the unwanted side-effect of excessive censure but also fails to inform us as to why transcendental illusion is illusory. “Particular errors can be got rid of by censure, [but] their causes by criticism” (A711/B739). Moreover, Kant thinks that his philosophy allows for a potential sublimation of any desire to speculate about transcendent things, such as God and the immortal soul. It does this by showing, first, that our pure concepts enable a kind of indeterminate thinking about transcendent things, and, second, that we can have no theoretical knowledge that such things do not exist. These findings, Kant thinks, secure our right to practical thinking, as opposed to theoretical knowing, about reason’s most important objects (Bxviii–xxii, A254–56/B310–12, A744–46/B772–74). This suggests that reason’s desire to speculate about such things can be redirected toward practical thinking, and so transmuted into a desire that we can actually satisfy. 188A159/B198, A180/B233. 189A198/B232. 190A213/B260, my emphasis.

141 that precedes all experience is certain, for what is more certain than experience.”191 Again, “sound common sense will always assert its rights in this domain.”192 A skeptical empiricist may adopt a methodology on which it is hard to see how we could know any such principle, or may deny that we ever can. In doing so, she erects an obstacle to understanding how she can know what she can, in experience, easily recognize to be true. If she does, on some level, recognize the truth of such principles in her employment of them in experience, as Kant thinks we “easily” do, she could become conscious of a tension within her thought. Her consciousness of this tension could spur a desire to clear away any obstacles to her understanding how she can know the principles she employs. Kant’s Deduction can satisfy that desire, by explaining how pure concepts can validly apply to the objects of experience, and so figure in substantive a priori knowledge. So the explanation it provides can appeal to the skeptical empiricist not only because it addresses our desire for metaphysical knowledge, about which she must remain curious, but because it can reconcile the skeptic with our common sense knowledge of the principles of the understanding.193 If what I say is correct, there is a limited sense in which Kant attributes an incoherence to the skeptic’s thinking after all. Kant thinks we can have synthetic a priori knowledge of principles which the skeptic doubts, and thinks that the skeptic remains in a position to recognize that we have such knowledge in virtue of her necessary employment of these principles in experience. The Deduction plays a crucial role in explaining our possession of such knowledge, and is offered to the skeptical empiricist as a means to escape her skeptical empiricism. But it is important to see this offer for what it is—as an offer. An attempt to compel the skeptical empiricist to recognize her a priori knowledge of principles in virtue

191MM 29:805. See also MM 29:794. 192P 4:351. 193A similar point can be made concerning natural-scientiﬁc and mathematical knowledge, both of which Kant repeatedly claims we can recognize to be “actual” (B4–5; B14–18, B20, B127–28; P 4:279–80, 4:294– 95), despite the skeptical empiricist’s doubt about or denial of such synthetic a priori knowledge (B127–28, A760/B788, A765/B793; KpV 5:13, 5:51). Kant is careful to clarify that Hume viewed his skepticism as not extending to mathematical knowledge. But Kant thinks that Hume was able to maintain this only because he failed to recognize that mathematics is synthetic a priori. Ultimately, Kant thinks that Hume would have given up his skeptical empiricism because of the threat it poses to our mathematical knowledge if he had recognized mathematics’ synthetic a priori status (B4–5; B19–20; P 4:260; 4:272; KpV 5:13–14; 5:53–54). As recent interpreters have noted, Kant would probably be wrong to think this. See, for example, Thielke(2015). Regardless, these passages suggest that Kant thought that the skeptic’s being in a position to recognize our actual possession of synthetic a priori knowledge could play a role in moving her to give up her skepticism.

142 of the role they play in making experience possible will beg the question against her, if she has not yet grasped how we can have a priori knowledge of objects. But forcing such an explanation upon her will have little effect. For doing so is likely to put her in a defensive posture, in which she holds fast to her empiricism, and so judges the explanation to be of little value. But if the explanation remains a standing offer, it can appeal to the skeptic when her dissatisfaction with her own position periodically or inevitably creeps through the cracks, either because it thwarts restless reason’s aims or because it clashes with common reason’s insight.194 In sum, Kant has a rationale for thinking that the explanation he offers will appeal to the skeptic, rather than simply begging the question against her. On his view, skeptical empiricism stands on shaky ground, based only on contingent facts about previous philosophical failures. It denies a kind of knowledge that common human reason recognizes us to have. And all reasoners, including skeptics, must remain interested in the possibility of synthetic a priori knowledge. By leading to skepticism, empiricism urges us to abandon all attempts to satisfy this interest. For that reason, adopting skeptical empiricism might seem to remove the appeal of an explanation of synthetic a priori knowledge. But, for Kant, the opposite occurs. By stifling our necessary interests, and doing so on insufficient grounds, skeptical empiricism turns out to be necessarily unstable. And because the view is unstable, it cannot stably stand in the way of pursuing those interests. A skeptical empiricist cannot be completely satisfied with her skepticism. The more dissatisfied she feels, the more likely she will be to hear out Kant’s explanation, and grasp its appeal.

194The “despair” and “melancholy” which Hume feels in T 1.4.7 in response to the skeptical implications of his finding the “trivial” imagination at the root of all belief and reasoning can itself suggest Hume’s dissatisfaction with his own position (T 1.4.7.1 and 3, SBN 263, 265). It is true that Hume portrays himself as emerging from this gloom with returned “curiosity” and “ambition” (T 1.4.7.12–13, SBN 270–71). But we might wonder how stable this change in humor is. Hume does not obviously overturn the lines of thought which led him into despair in the first place. And until Hume has clearly addressed the sources of the despair, we can worry that the despair will return. Any continued consciousness of such despair would give Hume a reason to hear out Kant’s offer. Speculatively, Kant could perhaps have judged as much from reading Hamann’s 1771 translation of T 1.4.7. See n. 164 above. For readings of Hume on which he develops resources for resolving the sources of his despair, see Garrett(2002), Ch. 12; Schafer(2014). For concerns, see Qu(2014). For more skeptical readings on which Hume has no resolutions, see Thielke(2003); Waxman (2003), 266–79.

143 4.7 FURTHER OBSTACLES

A skeptic’s commitment to the experimental method presents an obstacle to his acceptance of Kant’s philosophical doctrines. As I have argued, Kant viewed his offer as capable of cir- cumventing this obstacle by appealing to the skeptic’s recalcitrant interest in the possibility of synthetic a priori knowledge. But empiricism is not the only obstacle which can arise. To better understand the appeal of Kant’s offer, it will help to consider some further obstacles to accepting it. Although the empiricist’s method presents a crucial obstacle, Kant has a richer picture of the ways in which his offer might fail to reach a skeptic, and of the ways in which some of these difficulties can be overcome. One way in which Kant’s offer may fail to reach the skeptic is through a failure to effectively communicate his explanation, or by succeeding only partially without eradicating obstructions and obscurities. The explanation is indeed difficult to read and understand. Kant seems to admit as much in the preface to the 1781 edition of the Critique, when he describes his book’s “dry, purely scholastic fashion” of writing, its “large[ness] in bulk, and its “unsuitability” for “popular consumption.”195 Again, in the Prolegomena, Kant explains that readers of his Critique are apt to skim around and so misread what scattered passages catch their attention, overlooking the book’s systematicity. Such readers are, in Kant’s eyes, unwilling to “expend [the necessary] effort on it, because the work is dry, because it is obscure, because it opposes all familiar concepts and is long-winded as well.”196 This estimation of the Critique was apparently shared by Kant’s contemporaries. As the author of the anonymous 1781 review in the GöttingenReview put it, “the work. . . always exercises the understanding of its reader if not always instructing it, often strains the attention to exhaustion.”197 Perhaps no part of the Critique is more notoriously straining than the Deduction, which cost Kant “the greatest labor” to write, and has been described more recently as a “mystery,” a “jungle,” and “complex and elusive.”198 If the Deduction is

195Axviii. See also Bxxxiv. 196P 4:261. Cf. 4:262, 4:274, 4:277–78. 197The reviewer was Christian Garve. I quote Hatﬁeld’s translation from Kant(1783/2004), 201. 198Axvi; Strawson(1966/2002), 85; Bennett(1966), 100; Henrich(1969), 64, respectively. In a 1781 letter to Johann Erich Beiser, Kant complains that, in an attempt to get the ﬁrst edition Critique to the presses sooner, he failed to eradicate “certain stylistic infelicities and signs of haste as well as certain obscurities,” and hoped to make emendations after “the judgment of the public has called attention to the places that seem

144 indeed so difficult to read, it is thus likely to discourage readers without heroic attention and commitment. That itself could explain why many skeptics never found the way forward that Kant took himself to have seen and offered to them. Another source of obstacles is the various moods and attitudes which often accompany skepticism. Some of these encourage skeptics to become obstinate, poor listeners, even to the friendliest and most digestible offers. A skeptic may, for instance, find a twisted contentment in despair or self-pity and so be left cold by opportunities to emerge from it, like the slothful unrequited lover who enjoys his dissatisfaction. Again, a skeptic may take pride in her refusal to listen to those outside her ‘in’-group, like punks or dissenters proud of flouting the status quo. The promise of literary fame or professional prestige, too, may motivate a skeptic to concoct clever and novel arguments for his own position—a position which invites attack, yet appears impervious. In such cases, pursuit of recognition may pacify our restless reason for a while. For Kant, “personal vanity” often leads dogmatists to proceed with their metaphysical assertions, despite concealed doubts.199 But the same vanity can underlie a dogmatically inflexible rejection of all metaphysical questions, even despite the fickleness of skeptical doubt. This can seem to be how it is with Hume, who called his hope for “literary fame” his “ruling passion,”200 and whose palpable delight in his own ‘science of man,’ so ingeniously constructed from such meagre resources, drips from nearly every page of his Treatise. Such pseudo-satisfactions, and the personalities and moods which tend toward them, may postpone the skeptic’s natural urge to have metaphysical questions answered, or else dismissed on principle. The explanation whose core Kant offers in the Deduction may very well fail to wrest a skeptic out of these personalities and moods. But it is extreme to see this as a flaw. to need them” (C 10:272–73). Perhaps the Deduction was one of these “places,” as it received a complete overhaul in the second edition. But, as the citations of more recent interpreters attest, many continue to find the second edition Deduction to be obscure. 199A749/B777; cf. A10n; B256/A217; A423/B451; A757/785; A769/B797. 200E MOL 21, Mil xl–xli. It is rarely acknowledged that Hume mentions his passion for literary fame within a discussion of his own steady and sanguine temperament. He says: “I was, I say, a man of mild disposition, of command of temper, of an open, social, and cheerful humour, capable of attachment, but little susceptible of enmity, and of great moderation in all my passions. Even my love of literary fame, my ruling passion, never soured my temper, notwithstanding my frequent disappointments.” If we are to believe Hume’s own account, this ruling passion was not so strong as to shake his “command of temper” and ruin his openness and sociability. That suggests that his passion for fame would not have posed an obstacle to his listening to and patiently entertaining the alternative viewpoints of others.

145 Probably no piece of philosophy or rhetoric can appeal to every uncooperative character. The Deduction still offers skeptical empiricists a way out of their predicament, whether or not they are consistently open to it. Still, it is worth briefly addressing the worry that Kant viewed skeptics as having characters or moods particularly poorly suited for their going along with his philosophy. This worry can arise by focusing on some of the passages where Kant gives somewhat more negative portrayals of skeptics. For example, Kant’s first mention of skeptics in the A Preface describes them as “a species of nomads, despising all settled life, [who] broke up from time to time all civil society.”201 This attribution of unruly scorn paints the skeptic in an unflattering light, even in comparison to the warring dogmatists. And it seems to be reinforced elsewhere, for example, when Kant says that skeptics “would feign make show of having a philosophical justification for their contemptuous dislike of all enquiries of [a speculative] kind.”202 Here, we have both dishonesty and contempt. While such traits are generally uncooperative, they might seem to create a particular resistance to entertaining Kant’s philosophy. The skeptic’s dislike of metaphysics begins in a frustration with the vanity of dogmatic boasting, and the endless feuding this leads to. And then Kant comes along, claiming to have solved all metaphysical problems, and to have achieved complete knowledge of reason itself and of pure thinking. It is no wonder that Kant “fanc[ies] that he detect[s] in the face of the reader an expression of indignation, mingled with contempt, at pretensions seemingly so arrogant and vain-glorious.”203 Skeptics in particular seem primed to view Kant as just another dogmatic boaster. But to expect such hostility from the skeptic, Kant tells us, is to view his character only in light of his reactions to the dogmatist, and to overshadow his yearning for peace. Referring to the skeptical method of opposing dogmatic claims, Kant says:

To set reason at variance with itself, to supply it with weapons on both sides, and then look on, quietly and scoﬃngly, at the ﬁerce struggle, is not, from the dogmatic point of view, a seemly spectacle, but appears to suggest a mischievous and malevolent disposition. If, however, we consider the invincible obstinacy and the boastfulness of those who argue dogmatically, and who refuse to allow their claims to be moderated by any criticism, there

201Aix. 202A757/785. 203Axiii–xiv.

146 is really no other available course of action than to set against the boasting of the one side the no less justiﬁed boasting of the other, in the hope that the resistance thus oﬀered to reason may at least serve to disconcert it, to awaken some doubts as to its pretensions, and to make it willing to give a hearing to criticism.204

What to a dogmatist appears as contempt Kant charitably reads as blamelessly inevitable and even hopeful. It is a call to criticism, presumably only hazily envisioned, yet nonetheless recognized as a possible end to the warring. Though certainly capable of stalling and stewing, the skeptic has, for Kant, a halcyon forecast. That is perhaps why, in a 1768 letter to Johann Gottfried Herder in Riga, Kant praises Hume, his “most ingenious” skeptic, as a model of philosophic tranquility. Eagerly anticipating Herder’s own maturation into such a peace of mind, Kant says:

In the early unfolding of your talents I foresee with divers pleasures the time when your fruitful spirit, no longer so sorely driven by the warm impulse of youthful feeling, attains that serenity which is peaceful yet full of feeling and is the contemplative life of the philosopher, just the opposite of that dreamed of by the mystics. From what I know of you, I conﬁdently look forward to this epoch of your genius, of all states of mind the most ad- vantageous to its possessor and to the world, one in which Montaigne occupies the lowest place and Hume so far as I know the highest.205

If Hume embodied this tranquility in virtue of his exemplary skepticism, then Kant would have reason to trust that skeptics could be motivated to listen to his explanations because of his promises of a restful era of peace. And, beyond that, he clariﬁes that the seeming arro- gance of his complete system of metaphysics is based in a mature humility. His “pretensions,” he says,

are incomparably more moderate than the claims of all those writers who on the lines of the usual programme profess to prove the simple nature of the soul or the necessity of a ﬁrst beginning of the world. For while such writers pledge themselves to extend human knowledge beyond all limits of possible experience, I humbly confess that this is entirely beyond my power.206

204A757/785. 205C 10:73–74. Compare n. 200 above. Here, I use the translation given in Cassirer(1981), 85, rather than Zweig’s in Kant(1999), 94, as I believe the former correctly interprets the “state of mind” [ Gem¨utsverfassung] which Kant attributes to Hume to be “serenity” [Ruhe]. Zweig opts for “conﬁdence” [Zuversicht] instead. See Kant(1999), 96n5 for a discussion of an ambiguity in the last sentence. We may presume that Kant is here speaking of the mature Hume, given that he wrote this letter before the younger, sometimes despairing, sometimes warmly optimistic Hume of T 1.4.7 had appeared in a German translation. On Kant’s access to Hume’s texts, see n. 164 above. 206Axiv.

147 With this profession of modesty and opposition to dogmatism, Kant hopes to win the skeptic over to his side. It may indeed suffice to inspire in a skeptic who feels hope for peace, or dissatisfaction with her own skepticism, “the patience and impartiality of a judge” that Kant asks of all his readers.207 And a little patience may allow the skeptic to look beyond her empiricism and take to heart Kant’s explanation of our knowledge, and to finally address her reason’s drive for knowledge independent of experience. Like the difficulties in Kant’s writing, skepticism’s characteristically uncooperative moods constitute myriad obstacles to his offer. But, if Kant is right, some of these moods are less characteristic of skepticism than it might appear. Hope and cooperation have their place in the skeptic’s personality as well, and can temper the obstinacy of her more uncooperative moods. Though Kant’s offer may not reach all skeptics, in all moods, many will find it a welcome antidote to their despair, and a promising path to peace.

4.8 CONCLUSION

Many philosophers invoke Kant’s name and spirit when attempting ambitious refutations of skepticism. They model their arguments on the apparent form of the deduction, thinking this will imbue them with compelling force.208 Perhaps this strategy can succeed. But we must not let it overshadow a less aggressive, more cordial response to skepticism. This is to offer the skeptic a way of looking at things that she overlooks—one which illustrates or explains the possibility of what she doubts or denies. I have argued that we can find this friendly sort of offer in Kant’s Deduction. And I think that if we hold Kant’s offer in mind, we will be able to see other responses to skepticism along roughly the same lines. This can offer us an easily overlooked strategy for responding to, and curing, skepticism. This is the same explanatory strategy that might resolve a doubt about the possibility of a steam locomotive’s moving along railroad tracks. Or to take a less dated example, imagine attempting a mathematical proof of a so far unproven conjecture. After twelve attempts you

207Axxi. 208See Strawson(1959), 62–63; Putnam(1981), 16; Korsgaard(2009), 32–33.

148 find you cannot get it to work. You may seem to have good grounds to stop trying, based on your failures so far. You might also conclude with some justification that you do not know the conclusion of the proof. But now someone comes along and says she has completed the proof. You should not respond by saying, “That’s not possible! We’ve failed so many times.” You should instead look at her proof. If you see that it succeeds, you may come to know what you had decided could not be known. The example of a mathematical proof brings out one way in which one might plausibly disagree with Kant. His insistence that ignorance should always prompt continued inquiry unless one has found the ignorance to be necessary seems extreme, even for cases of a priori reasoning. We may be right to give up after twelve attempts. The empiricist may likewise have been right to give up on the possibility of synthetic a priori knowledge after observing the repeated failures of previous metaphysicians. I doubt that this latter point is true, but I have to admit that it is at least initially plausible. Nevertheless, he would be wrong not to listen to the explanation of a possibility someone thinks he has overlooked. And he would be dogmatic to reject the explanation on the grounds of previous failures. In the case of the mathematical proof, listening to the explanation can lead to genuinely new knowledge. In Kant’s case, the explanation can lead the skeptic to recognize that he has had the knowledge in question all along, without ever needing a proof. In both cases, we can see how the explanation can naturally, and rightly, appeal to the skeptic, at least when he is not at his most stubborn or despairing. If Kant is right, this appeal is even stronger in the case of the skeptical empiricist than in that of the mathematical proof. The skeptical empiricist, on Kant’s view, has a necessary interest in the knowledge in question, which no despair can fully eradicate. And she already possesses, albeit in a hazy way, the knowledge she has come to doubt. This can make its clarification more likely to appeal to her, and her doubts all the more unstable. The explanation can seem to do little if it does not answer the quid juris, showing the validity or legitimacy of applying the categories to the objects of experience. Bernhard Thöle expresses this concern when he says: “The Transcendental Deduction cannot constrain itself to an explanation of the manner in which pure concepts of the understanding can apply to objects: the objective validity of the pure concepts of the understanding must also be

149 proven.”209 I think this is wrong in two ways. First, contrary to both Thöleand the oppositional reading, a proof, or a non-question-begging refutation, is not the only way to answer the question of right. Second, it may be that nothing further is needed at all; the explanation itself answers the question of right. If Henrich is correct to put the question of right in the skeptic’s mouth, then we can see Kant’s explanation in the Deduction as answering that question. It does this by explaining a possibility she overlooks, thereby removing the concern which made the question seem necessary. In this regard, I agree with Hatfield that the Deduction can interest us in virtue of what it tells us about “the place of theories of the cognitive faculties in the theory of knowledge, or [of] the various relations that existed between epistemology and psychology.”210 But, unlike Hatfield and other proponents of the inheritor reading, I do not think we must view the Deduction as making no attempt to convince Hume that we can have synthetic a priori knowledge in order for it to interest us in this way. On the contrary, the Deduction’s own interest in the workings of the cognitive faculties is part and parcel of its strategy for convincing skeptical empiricists. More generally, one way to answer a question of right is to show the questioner that her doubt is an artifact of neglecting an alternative picture of the mind which is then explained. That is, I have argued, what Kant does in his Deduction. My argument in this chapter has been primarily exegetical. I have not offered a detailed defense of Kant’s explanation, either on its own terms or as a response to skepticism. Instead, I have laid the groundwork for that defense. Coming to a different understanding of Kant’s project in the Deduction is likely to open up new avenues for defending that project. In describing the project, I have also described a rationale—which I argued is Kant’s—for thinking that the skeptic is likely to be at least partly dissatisfied with his own skepticism, in a way that will make it easier to reach him. The Deduction’s friendly offer, together with the instability of the skeptic’s position, provide reasons for thinking that there is a viable strategy for changing skeptics’ minds. Those reasons become even stronger when we reflect on the relations between the chapters in this dissertation. Let us now turn to those relations.

209Thöle(1991), 35, my translation of “ Die transzendentale Deduktion kann sich. . . nicht auf eine Erklärung der Art, wie sich reine Verstandesbegriffe auf Gegenständebeziehen können,beschränken: die objektive Gültigkeitder reinen Verstandesbegriffe muß auch bewiesen werden.” 210Hatfield(2001), 207.

150 5.0 CONCLUSION TO THE DISSERTATION

I have presented these chapters as self-standing essays. Chapter1 motivated the need for cure as a response to skepticism. Each of the following three chapters emphasized a part or element of a viable cure. Chapter2 argued that skepticism about the external world is groundless: that typical skeptical arguments assume a principle I called No Guarantee, on which perception can never put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be. Chapter3 presented and defended Hume’s view that skepticism is primarily a temperament or mood, which can be healthy in moderation but a ‘malady’ when excessive or dominant. Chapter4 argued that Kant aims to change a skeptic’s mind by offering an explanation of a possibility she overlooked. Though it is useful to consider each chapter on its own, they are more effective when they are combined. For one thing, the dominant theme of each chapter echoes in the other chapters. Noticing these echoes helps us see how the themes fit together. Chapter2’s revealing the groundlessness of skepticism parallels Hume’s tracing skepticism to temperamental excess more than to argument (Chapter3), and Kant’s discussion of the insufficient and ultimately shaky grounds underlying the skeptic’s empiricist methodology (Chapter4). Chapter 3’s argument that Hume views skepticism as primarily a mood or temperament echoes in my pointing to feeling as a natural source of the skeptic’s groundlessly adopting No Guaran- tee (Chapter2), and in Kant’s characterization of skeptical empiricism as a kind of despair (Chapter4). Chapter4’s reading of Kant as offering the skeptic an alternative parallels my offering the skeptic an alternative conception of experience on which it can provide us with guarantees (Chapter2), and Hume’s offering a conception of skepticism in humoral terms as a way to guide the management and treatment of pathological skepticism and other imbalances (Chapter3). It is not surprising that these themes echo across the chapters when we see that the themes, and the chapters, complement one another, combining to form a cure for skepticism. By tracing skepticism to its source in groundless intuition and feeling, Chapter2 shows the appropriateness of Chapter3’s treatment of skepticism, which focuses on the moderation

151 of temperaments and moods. In return, Chapter3’s treatment of skepticism presents a way to moderate the strength of the felt intuition which makes skeptical arguments appear plausible. Together, Chapters2 and3 connect skepticism to feeling and temperament, and show that skepticism’s proper treatment has a lot to gain from taking into account the skeptic’s characteristic moods. In some cases, perhaps only a shift in mood will drain intuitions like No Guarantee of some of the force with which the skeptic is apt to feel them. Chapters3 and4 together shed light on both the appeal of, and the obstacles to, a skeptic’s entertaining an alternative explanation of our knowledge which he has overlooked. This is because the melancholy or despair which both Hume and Kant find central to skepticism is a complex emotion, which can be both an obstacle to and a motive for cure. On the one hand, despair reinforces itself. Its overabundance makes a person hopeless, and liable to give up on understanding the possibility of the knowledge he doubts. It also makes such a person unsociable and unwilling to accept help. On the other hand, the unpleasantness of despair drives the skeptic to escape it. Insofar as the offer of an overlooked alternative promises to explain the knowledge which the skeptic despairs of understanding, it can appeal as a means to remove that despair. If the skeptic finds the explanation is novel, sound, and enlighten- ing, it can succeed. In this way, Chapters3 and4 together illuminate this dual structure of skeptical despair. They thus explain why the skeptic is, on the one hand, stubborn and recalcitrant, and, on the other, open to—even calling for—help. Both sides of this structure teach us about the viability and limitations of a cure. Coming full circle, Chapters4 and2 combine to diagnose the skeptic’s groundless intuition as a lack of the understanding which the cure’s alternative explanation of our knowledge offers. In the first instance, a missed alternative makes the groundless intuition much more robust. To a person who overlooks a satisfying explanation of the possibility of our possessing some kind of knowledge, it can feel plausible, even compulsory, to conclude that we cannot have that kind of knowledge. With that feeling in place, arguments to the effect that we cannot know such things seem compelling, and become a further barrier to the skeptic’s grasping the overlooked possibility. But these arguments can be disarmed by demonstrating that they rely on a groundless intuition. Toppling the arguments then makes it more natural for the skeptic to entertain the alternative, and through it come to grips with our knowing

152 what we ordinarily take ourselves to know. Chapters2 and4 in particular may combine in subtler ways. Plausibly, Kant’s explanation of how certain central concepts figure in our empirical knowledge may provide a deeper explanation of how perception guarantees the truth of what we perceive. This deeper explanation may not always be necessary to change the mind of a skeptic about perceptual knowledge. For example, after coming to see No Guarantee as groundless, a skeptic about the external world may be able to accept the idea that perception can provide us with guarantees on ordinary grounds. But to a skeptical empiricist who finds No Guarantee plausible because she cannot grasp how we can come, through experience, to have knowledge of objects standing in necessary causal connections, a story about how concepts such as ‘cause’ make such experience possible could play a central role in her seeing how perception could provide us with a guarantee of the truth of what we perceive. It could do this in part by giving the skeptic a way of understanding perception as putting us in touch with objects, by way of categorially unified intuitions rather than only fleeting sense impressions.1 We can combine the insights in each chapter into a general prescription for cure: Show that the considerations which lead to skepticism rely on an assumption that is unsupported, and so may be left behind; if needed, treat any moods which may remain in the way of the skeptic’s natural interest in understanding the possibility of our knowledge; provide an explanation of that possibility which the skeptic has so far failed to acknowledge or comprehend. In saying that the chapters combine to give a cure, I do not mean that no chapter gives a complete cure on its own. In fact, I think that Chapters2 and4 can each reach a skeptic and change her mind without needing to bring in material from the other chapters. Chapter

1Kant’s explanation does so by offering an account of an empirical intuition—in other words, an account of an element of experience through which a subject has an object sensibly given to her. Kant’s account tells us that we are given objects only through intuitions which are unified by the forms of our thought—namely, by the pure concepts, or categories. According to him, the application of these categories makes it possible for objects to be sensibly given to us, because it makes them thinkable. Insofar as the given objects are, through thought, made present to our consciousness, we have a guarantee of their existence. And, by the same token, we have a guarantee that they are the ways that we think them to be by applying the categories. A skeptic who does not have the idea that the forms of thought operate within our intuition may give up on the idea that objects can ever be present to our consciousness. And so they may find No Guarantee compelling. But this groundless presupposition can be avoided by a story which explains the role of thought in our perception, or, more generally, in all our intuition.

153 2 argues that skeptical arguments rely on No Guarantee; shows this is a groundless intuition; and offers epistemic disjunctivism as an alternative conception of perception and its relation to knowledge. Chapter4’s Kant sees skeptical empiricism as based on insufficient grounds; views this skepticism to be a despair as regards reason’s most important aims; and offers an explanation of the application of pure concepts as a way of making sense of empirical knowledge. Nonetheless, as we saw, the individually complete cures bolster one another in further ways. Cocktails of medicines which are themselves already complete cures sometimes make for more effective cures. Chapters2 and4, together with the discussion of moods and temperaments in Chapter3, can be more convincing, better illuminate our knowledge, and reach more people. Their combination better serves the humanitarian, educational, and preventive aims I described at the end of §1.5. These are the aims of cure. Along the way, we can also learn from listening to the skeptic—rather than from curing her—how to moderate other kinds of excesses related to belief, such as dogmatism (Chapter 3). We can even be shaken from our own dogmatism by skeptical arguments, which call us to better understand the limits of our knowledge (Chapters3 and4). Although the danger of falling into skeptical excess will always be with us, a moderate degree of the very same disposition which leads us into this danger is something each of us should strive for. As Hume emphasized (Chapter3), this disposition to wonder, and reason, and inquire is at the core of who we are.2 This is why the danger of excessive skepticism will always be with us. We might worry that the cure I have offered does not succeed because an explanation of the possibility of our knowledge does not show that we actually have that knowledge. Isn’t there still a further step to be made here to support the claims to the knowledge we ordinarily take ourselves to have? This further step can seem to be what the skeptic demands. But as we saw in Chapter2, a call for a further step is groundless. Allowing us to see how perception can give guarantees allows us to “actualize” this knowledge on the basis of perception. The support for common-sense knowledge is perception. To think that an instance of perceptual knowledge requires something further than what perception alone can provide is to think

2As Aristotle says, “All men by nature desire to know” (Metaphysics, Book I, Ch. 1, 980a20). Translation from Barnes(1984), 1552.

154 that perception alone can never guarantee that things are as we seem to perceive them to be. It is to fall back into No Guarantee. Relatedly, we might worry that even if No Guarantee is groundless, the alternative to it is equally groundless. Why accept a theory on which perception can guarantee that things are as we seem to perceive them to be? Without some compelling reason to accept the alternative, it can seem that we will be stuck at an impasse with the skeptic, and that the conversation will go on indefinitely without either side budging. But there is no impasse. “Grounds for doubt are lacking!”3 When we lack grounds to doubt something, the ordinary ways we think about that thing are the starting points of a theory. These starting points provide the grounds for the theory. We can recognize that the starting points are proper grounds once we have cleared away the skepticism which seems to shake them. The skeptic who does not yet recognize our knowledge may not find the grounds compelling. But we should understand this wariness to be an effect of her doubt, rather than a justification for it. Removing the root of her skepticism can allow her to recognize the grounds for the theory. We can thus reply to these objections about explaining a mere possibility, or offering a groundless alternative. But these objections can also be misleading in another way. A focus on them can portray the skeptic as already aware of the alternative being offered, without seeing its appeal. If such skeptics exist, the best response to them may be the argument I made in Chapter2. But we have also seen some reasons to think that this is not an accurate portrayal of many skeptics. Although some may appear stubborn, skeptics are typically perplexed. They typically want to understand our knowledge. Even their despair about understanding our knowledge is a motivation for considering alternatives which may allow them to emerge from their skepticism. These motivations, too, make an impasse less likely. Some skeptics may consider and accept a successful alternative explanation of our knowledge even before their apparent grounds for doubt have been removed. Accepting the alternative then immediately removes those apparent grounds. In other cases, those apparent grounds may be removed first. The skeptic will then be all the more likely to consider and accept an alternative. That alternative has good grounds, and appeals to her even without them. We can see this in Kant’s discussion of the skeptical empiricist in Chapter4. Kant

3Wittgenstein(1969/1972), §4: “Zum Zweifeln fehlen mir die Gr¨unde!”

155 portrays the skeptical empiricist as wanting to escape her situation, and respond to innate drives of reason which push her to return to an enterprise which she has abandoned out of frustration. Although I find it quite plausible that we have these drives, we do not need to attribute them to ourselves in order to see that Kant was right to find within skepticism a kind of dissatisfaction and a related, though perhaps submerged, aspiration to understand human knowledge. In some cases, that aspiration plays a significant role in the generation of skeptical arguments, as part of an assessment of our knowledge. In those cases, there is little reason to think that the desire must have been completely eradicated, or cannot be coaxed out by a hope for cure. But more generally, the tension a skeptic must feel with our ordinary picture of our knowledge, and indeed with that knowledge itself, is already significant. It is psychologically realistic to think that such a tension can generate a desire to understand our knowledge—a desire which can coexist with and outlive any professed denials of such knowledge. In the dilemma in Chapter1, the perplexed person was found to be similar to the skeptic, and similarly in need of a cure. Conversely, we can now see that the skeptic is much like the perplexed person. Neither party is whole-hearted. Both have a tether to common-sense, and a motivation to look for a resolution to their perplexity about our knowledge. Understanding this point both buttresses the dilemma and indicates the viability of a cure. In the Preface, I proposed to view skepticism as more than just an overt denial of our knowledge. It can also be, I claimed, a doubt or bewilderment or discomfort or alienation or frustration that one can feel with regard to our knowledge—a perplexity at the core of many varieties of skepticism. The chapters of this dissertation have both used and defended this conception. They have defended it by tracing skepticism to a root cause in feeling, by characterizing and analyzing skepticism’s characteristic moods, and by recognizing that an explanation of our knowledge can address the skeptic. They have used it to develop a more effective cure, combining a refutation of apparent grounds, a cultivation of opposing moods, and an offer of an explanation of our knowledge that appeals to someone in an uncomfortable state of frustration. If I am right, we can and should talk to skeptics. We need not be quiet or dogmatic in their presence. We can show them the groundlessness of their views, instead of conceding the

156 groundlessness of ours. We can learn from them in the process, as Hume and Kant do. And we can change their minds by oﬀering a compelling, alternative conception of our connection to the world.

157 APPENDIX A: CHAPTER SUMMARIES

Chapter 1: “Cure and Prevention”

Many philosophers now think that there is no use in trying to convince a skeptic that we know what she claims we cannot know. Having given up on trying to cure the skeptic, they opt instead for a merely preventive response. This kind of response aims only to show non-skeptics that skeptical arguments fail, even if this begs the question against the skeptic. Against this trend, I argue that merely preventive responses are likely to fail. For a response to skepticism to succeed, I argue, it must inquire into why skeptical arguments appear plausible. The merely preventive response then faces a dilemma: Either the inquiry reveals that the responder is himself in need of a cure or else the inquiry constitutes a cure.

Chapter 2: “The Groundlessness of Skepticism”

I oﬀer a three-part cure for skepticism about the external world. First, I show that prominent arguments for this kind of skepticism rely on a single, shared idea. This is the idea that perception can never put us in touch with the world in a way that guarantees that things are as we seem to perceive them to be. Second, I show that the best available arguments for this idea are question-begging. It follows that, unless a better argument can be given, the idea should be treated as a baseless intuition. Third, I argue that the presentation of an alternative view about what perception can guarantee—a view sometimes called ‘epistemic disjunctivism’—completes the cure. Once the external world skeptic is made to see that her skepticism is groundless, she can accept the alternative view on ordinary grounds. Since this view oﬀers a straightforward way of making sense of perceptual knowledge, the skeptic’s acceptance of it cures her.

158 Chapter3: “The Humors in Hume’s Skepticism”

In the conclusion to the first book of A Treatise of Human Nature, Hume’s skeptical reflections have plunged him into melancholy. He then proceeds through a complex series of stages, resulting in renewed interest in philosophy. Interpreters have struggled to explain the connection between the stages. I argue that Hume’s repeated invocation of the four humors of ancient and medieval medicine explains the succession, and sheds a new light on the significance of skepticism. The humoral context not only reveals that Hume conceives of skepticism primarily as a temperament, not a philosophical view or system. It also resolves a puzzle about how Hume can view skepticism as both an illness and a cure. The skeptical temperament can, depending on its degree of predominance, either contribute to or upset the balance of temperaments required for proper mental functioning.

Chapter4: “Kant’s Oﬀer to the Skeptical Empiricist”

There is little consensus about whether Kant intends his Critique of Pure Reason to change a skeptic’s mind. I challenge a common assumption made by both sides of the debate. This is the thought that Kant can only convince a skeptic if he does not beg the question against her. That, I argue, is not how Kant sees things. On Kant’s view, skeptical empiricism is an inherently unstable and unsatisfying position, which skeptics cannot help wanting to escape. Kant’s Critique, and especially its Transcendental Deduction, oﬀers an appealing means of escape, by explaining a possible relation of the mind to the objects of knowledge which the skeptic has overlooked.

Chapter5: “Conclusion to the Dissertation”

I explain that Chapters2,3, and4 emphasize diﬀerent elements of the cure for skepticism motivated in Chapter1, and argue that they combine to mutually reinforce the viability of such a cure.

159 APPENDIX B: NOTE ON CITATIONS OF HUME AND KANT

Hume

All citations of A Treatise of Human Nature text will be marked ‘T,’ followed by the book, part, section, paragraph, and page number of the Selby-Bigge Nidditch edition, respectively; ‘A’ is used to refer to Hume’s Advertisement to the Treatise; the text for both is derived from the Norton & Norton Clarendon Press edition (Hume, 1739/2007). Citations of An Enquiry Concerning Human Understanding are marked ‘EHU’ followed by section and paragraph numbers, and the page number of the Selby-Bigge Nidditch edition; the text is derived from the Beauchamp Clarendon Press edition (Hume, 1748/2000). Citations of Hume’s Essays, Moral, Political, and Literary use ‘E’, followed by the standard essay ti-tle abbreviations listed on http://www.davidhume.org, paragraph number, and then page number of the Miller edition (Hume, 1758/1987). Citations of the History of England use ‘H’ followed by volume, chapter, and paragraph number; the text is derived from the Lib-erty Fund Edition (Hume, 1778/1983). Citations of Dialogues Concerning Natural Religion use ‘DNR,’ followed by the section and paragraph number, and then the page number of Kemp Smith’s edition (Hume, 1779/1947). Citations of Dissertation of the Passions and The Natural History of Religion use ‘DP’ and ‘NHR,’ respectively, followed by the section, paragraph, and page number of the Beauchamp edition (Hume, 1757/2007). Citations of Hume’s letters use ‘HL’, followed by the letter number given in Greig (1932/2011), and the page number of that edition.

Kant

All citations of Kant’s Critique of Pure Reason use the standard A/B pagination to refer to pages of the 1781 (A) and 1787 (B) editions. I use Norman Kemp Smith’s 1929

160 translation. All other citations of Kant use the standard Akademie Ausgabe page numbers. I use the following abbreviations for Kant’s writings: ‘VKK’ for “Essay on the maladies of the head,” ‘P’ for Prolegomena to Any Future Metaphysics, ‘KpV’ for Critique of Practical Reason, ‘KU’ for Critique of the Power of Judgment, ‘ UE’¨ for “On a discovery whereby any new critique of pure reason is to be made superfluous by an older one,” ‘FM’ for “What real progress has metaphysics made in Germany since the time of Leibniz and Wolff?”, ‘TeFP’ for “Proclamation of the imminent conclusion of a treaty of perpetual peace in philosophy,” ‘Anth’ for Anthropology from a Pragmatic Point of View, and ‘C’ for Kant’s correspondences. The translations used are Holly Wilson’s (2007), Gary Hatfield’s (2004), Mary Gregor’s (1999), Paul Guyer & Eric Matthews’ (2000), Henry Allison & Peter Heath’s (2002) for the three essays, Robert Louden’s (2007), and Arnulf Zweig’s (1999), respectively. Citations of student-recorded notes on Kant’s lectures use the following abbreviations: ‘BL’ for Blomberg Logic, ‘DWL’ for Dohna-Wundlacken Logic, ‘JL’ for Jäsche Logic, and ‘MM’ for Metaphysics Mrongovius. The translations used for the logic lectures are J. Michael Young’s (1992), while the metaphysics lectures use Karl Ameriks & Steve Naragon’s (1999). Following Kemp Smith, I translate ‘Erkenntnis’ as ‘knowledge’ in all quotations of Kant, rather than ‘cognition,’ which is the Latin term Kant connects with his German expression. Any further deviations from the cited editions are discussed in footnotes. I prefer ‘knowledge,’ in part because of its familiarity as a non-technical, everyday term, and in part because it more obviously refers to a kind of achievement in virtue of which we are related to objective reality. Since I do not discuss Kant’s notion of ‘Wissen’ below, any ambiguity with that notion presents no problem. For recent discussions of the differences and relations between these notions, see Watkins and Willaschek (2020) and Tolley (2020).

Other

As per APA style, when two dates appear within a given citation (separated by a slash), the ﬁrst refers to the year of the ﬁrst publication of the relevant text, while the second refers to the year of the publication of the cited version. In the bibliography, some especially lengthy titles of early modern texts have been curtailed with ellipses to save on space.

161 Hume Kant

A Treatise of Human Nature T VKK “Essay on the maladies. . . head” Abstract to the Treatise Ab A/B Critique of Pure Reason (1781/7) Enquiry. . . Human Understanding EHU P Prolegomena to. . . Metaphysics Essays, Moral, Polit., & Literary E KpV Critique of Practical Reason “The Sceptic” E Sc KU Critique of. . . Judgment “Of Superstition and Enthusiasm” E SE UE¨ “On a discovery. . . ” History of England H FM “What real progress. . . ?” Dialogues Conc. Natural Religion DNR TeFP “Proclamation. . . Peace in Phil.” Dissertation of the Passions DP Anth Anthropology. . . Point of View Natural History of Religion NHR C Correspondences “My Own Life” MOL BL Blomberg Logic Hume’s Letters (ed. Greig) HL DWL Dohna-Wundlacken Logic JL J¨ascheLogic MM Metaphysics Mrongovius

Figure 2: Abbreviations for Hume and Kant’s writings

162 BIBLIOGRAPHY

Ainslie, Donald C. 2015. Hume’s True Scepticism. Oxford: Oxford University Press.

Alanen, Lilli. 2008. “The Powers and Mechanisms of the Passions.” In Traiger (2008), pp. 179–98.

Allais, Lucy. 2010. “Transcendental Idealism and the Transcendental Deduction.” In Schult- ing and Verburgt (2010), pp. 91–107.

Allison, Henry E. 1983/2004. Kant’s Transcendental Idealism. New Haven, CT: Yale Uni- versity Press.

———. 2008a. Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise. Oxford: Oxford University Press.

———. 2008b. ““Whatever Begins to Be Must Have a Cause of Existence”: Hume’s Analysis and Kant’s Response.” Philosophy and Phenomenological Research 76(3), pp. 525–46.

Ambrose, Alice. 1952. “Moore’s ‘Proof of an External World’.” In Schilpp (1952), pp. 395–418.

Ameriks, Karl. 1978. “Kant’s Transcendental Deduction as a Regressive Argument.” Kant- Studien 69(1-4), pp. 273–87.

———. 2003. Interpreting Kant’s Critiques. Oxford: Clarendon Press.

Anderson, Abraham. 1998. “On the Practical Foundation of Kant’s Response to Epistemic Skepticism.” Kant-Studien 89(2), pp. 145–66.

Austin, J.L. 1962/1964. Sense and Sensibilia, ed. G.J. Warnock. New York: Oxford Univer- sity Press.

Babb, Lawrence. 1951. The Elizabethan Malady: A Study of Melancholia in English Litera- ture from 1580 to 1642. East Lansing, MI: Michigan State University Press.

Baier, Annette C. 1991. A Progress of Sentiments: Reﬂections on Hume’s Treatise. Cam- bridge, MA: Harvard University Press.

Barnes, Jonathan, ed. 1984. The Complete Works of Aristotle: The Revised Oxford Trans- lation. Princeton, NJ: Princeton University Press.

Bennett, Jonathan. 1966. Kant’s Analytic. Cambridge: Cambridge University Press.

163 Bieri, Peter, Rolf-Peter Horstmann, and Lorenz Kr¨uger,eds. 1979. Transcendental Argu- ments and Science: Essays in Epistemology. Dordrecht: D. Reidel Publishing Company.

Boghossian, Paul. 2012. “Blind Rule-Following.” In Wright and Coliva(2012), pp. 27–48.

BonJour, Laurence. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press.

Bright, Timothie. 1586/1940. A Treatise of Melancholie: Containing the causes thereof, & reasons of the strange eﬀects it worketh in our minds and bodies. . . . New York: Columbia University Press.

Brook, Andrew. 1997. Kant and the Mind. Cambridge: Cambridge University Press.

Broughton, Janet. 2005. “Hume’s Voyage.” In Jenkins et al.(2005), pp. 174–91.

———. 2008. “Impressions and Ideas.” In Traiger(2008), pp. 43–58.

Brown, Theodore M. 1968. The Mechanical Philosophy and the “Animal Oeconomy”: A Study in the Development of English Physiology in the Seventeenth and Early Eighteenth Century. Ph.D. thesis, Princeton University.

Brueckner, Anthony. 1984. “Transcendental Arguments II.” Noˆus 18(2), pp. 197–225.

———. 1994. “The Structure of the Skeptical Argument.” Philosophy and Phenomenological Research 54(4), pp. 827–35.

———. 1996. “Modest Transcendental Arguments.” Philosophical Perspectives 10, pp. 265– 80.

Brueckner, Anthony and Gary Ebbs. 2012. Debating Self-Knowledge. Cambridge: Cambridge University Press.

Buckwalter, Wesley. 2017. “Epistemic Contextualism and Linguistic Behavior.” In Ichikawa (2017), pp. 44–56.

Burge, Tyler. 1993. “Content Preservation.” Philosophical Review 102(4), pp. 457–88.

———. 2005. “Disjunctivism and Perceptual Psychology.” Philosophical Topics 33(1), pp. 1–78.

———. 2011. “Disjunctivism Again.” Philosophical Explorations 14(1), pp. 43–80.

Burton, John Hill. 1846. Life and Correspondence of David Hume. Edinburgh: William Tait.

Burton, Robert. 1621/1964. The Anatomy of Melancholy. Everyman’s Library, London: Aldine Press.

164 Byrne, Alex. 2004. “How Hard Are the Sceptical Paradoxes?” Noˆus 38(2), pp. 299–325.

———. 2014. “McDowell and Wright on Anti-Scepticism, etc.” In Dodd and Zardini(2014), pp. 275–97.

Carter, J. Adam. 2012. “Recent Work on Moore’s Proof.” International Journal for the Study of Skepticism 2(2), pp. 115–44.

Cassam, Quassim. 1987. “Transcendental Arguments, Transcendental Synthesis and Tran- scendental Idealism.” Philosophical Quarterly 37(149), pp. 355–78.

Cassirer, Ernst. 1981. Kant’s Life and Thought. New Haven, CT: Yale University Press.

Chance, Brian A. 2011. “Sensibilism, Psychologism, and Kant’s Debt to Hume.” Kantian Review 16(3), pp. 325–49.

———. 2013. “Causal Powers, Hume’s Early German Critics, and Kant’s Response to Hume.” Kant-Studien 104(2), pp. 213–36.

Cheyne, George. 1724. An essay of health and long life. London: George Strahan.

———. 1733. The English malady: or, A treatise of nervous diseases of all kinds, as spleen, vapours, lowness of spirits, hypochondriacal, and hysterical distempers, etc. Lon- don: George Strahan.

Chignell, Andrew and Colin McLear. 2010. “Three Skeptics and the Critique: Review of Michael Forster’s Kant and Skepticism.” Philosophical Books 51(4), pp. 228–44.

Cicero, Marcus Tullius. 1877. Cicero’s Tusculan Disputations, ed. C.D. Yonge. New York: Harper & Brothers.

Cohen, Stewart. 1988. “How to Be a Fallibilist.” Philosophical Perspectives 2, pp. 91–123.

Comesa˜na,Juan and Peter Klein. 2019. “Skepticism.” In The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta, winter 2019 ed. Metaphysics Research Lab, Stanford University.

Dancy, Jonathan, Ernest Sosa, and Matthias Steup, eds. 2010. A Companion to Epistemol- ogy. Oxford: Blackwell. de Boer, Karin. 2019. “Kant’s Response to Hume’s Critique of Pure Reason.” Archiv f¨ur Geschichte der Philosophie 101(3), pp. 376–406.

DeRose, Keith. 1995. “Solving the Skeptical Problem.” Philosophical Review 104(1), pp. 1–52.

———. 1999. “Introduction: Responding to Skepticism.” In DeRose and Warﬁeld(1999), pp. 1–24.

165 ———. 2017. The Appearance of Ignorance: Knowledge, Skepticism, and Context, vol. 2. Oxford: Oxford University Press.

DeRose, Keith and Ted Warﬁeld, eds. 1999. Skepticism: A Contemporary Reader. Oxford: Oxford University Press.

Descartes, Ren´e.1641/1996. Meditations on First Philosophy: With Selections from the Objections and Replies. Cambridge: Cambridge University Press.

Dingwall, Helen. 1995. Physicians, Surgeons and Apothecaries: Medicine in Seventeenth- century Edinburgh. East Lothian: Tuckwell Press.

Dodd, Dylan and Elia Zardini, eds. 2014. Scepticism and Perceptual Justiﬁcation. Oxford: Oxford University Press.

Dretske, Fred. 1970. “Epistemic Operators.” Journal of Philosophy 67(24), pp. 1007–23.

Du Laurens, Andr´e.1594/1938. A Discourse of the Preservation of the Sight: of Melancholike Diseases; of Rheumes, and of Old Age. . . . London: London Shakespeare Association.

Dubos, Abb´eJean-Baptise. 1748. Critical reﬂections on poetry, painting and music. With an inquiry into the rise and progress of the theatrical entertainments of the ancients. . . . London: John Nourse.

Dyck, Corey W. 2011. “Kant’s Transcendental Deduction and the Ghosts of Descartes and Hume.” British Journal for the History of Philosophy 19(3), pp. 473–96.

Echeverri, Santiago. 2017. “How to Undercut Radical Skepticism.” Philosophical Stud- ies 174(5), pp. 1299–1321.

Edgar, Scott. 2010. “The Explanatory Structure of the Transcendental Deduction and a Cognitive Interpretation of the First Critique.” Canadian Journal of Philosophy 40(2), pp. 285–314.

Engstrom, Stephen. 1994. “The Transcendental Deduction and Skepticism.” Journal of the History of Philosophy 32(3), pp. 359–80.

Enoch, David and Joshua Schechter. 2008. “How Are Basic Belief-Forming Methods Justi- ﬁed?” Philosophy and Phenomenological Research 76(3), pp. 547–79.

Ertl, Wolfgang. 2002. “Hume’s Antinomy and Kant’s Critical Turn.” British Journal for the History of Philosophy 10(4), pp. 617–40.

Fogelin, Robert J. 1985. Hume’s Skepticism in the Treatise of Human Nature. London: Routledge & Kegan Paul.

———. 1993. “Hume’s scepticism.” In Norton(1993), pp. 90–116.

———. 2009. Hume’s Skeptical Crisis: A Textual Study. Oxford: Oxford University Press.

166 F¨orster,Eckart, ed. 1989. Kant’s Transcendental Deductions: The Three Critiques and the Opus Postumum. Palo Alto, CA: Stanford University Press.

Forster, Michael N. 2010. Kant and Skepticism. Princeton, NJ: Princeton University Press.

Frasca-Spada, Marina. 2003. “Belief and Animal Spirits in Hume’s Treatise.” Eighteenth- Century Thought 1(1), pp. 151–69.

French, Craig and Lee Walters. 2016. “The Invalidity of the Argument From Illusion.” American Philosophical Quarterly (4), pp. 357–64.

Garrett, Don. 2002. Cognition and Commitment in Hume’s Philosophy. Oxford: Oxford University Press.

———. 2008. “Hume’s Conclusion in “Conclusion of this Book”.” In Traiger(2008), pp. 151–75.

Gawlick, G¨unter and Lothar Kreimendahl. 1987. Hume in der Deutschen Aufkl¨arungUmrisse Einer Rezeptionsgeschichte. Stuttgart: Frommann-Holzboog.

Ginsborg, Hannah. 2011. “Primitive Normativity and Skepticism About Rules.” Journal of Philosophy 108(5), pp. 227–54.

Goldhaber, Charles. forthcoming-a. “The Humors in Hume’s Skepticism.” Ergo: An Open Access Journal of Philosophy.

———. forthcoming-b. “Does perceptual psychology rule out disjunctivism in the theory of perception?” Synthese. Online version published December 2019.

———. ms. “Transcendental Cures for Skepticism.” Unpublished manuscript.

Grant, Mark. 2000. Galen on Food and Diet. London: Routledge.

Grayling, A.C. 2010. “Transcendental Arguments.” In Dancy et al.(2010), pp. 768–71.

Greco, Daniel. 2012. “The Impossibility of Skepticism.” Philosophical Review 121(3), pp. 317–58.

Greco, John. 2007. Putting Skeptics in Their Place: The Nature of Skeptical Arguments and Their Role in Philosophical Inquiry. Cambridge: Cambridge University Press.

Greco, John, ed. 2008. The Oxford Handbook of Skepticism. Oxford: Oxford University Press.

Greenberg, Sean. 2008. “‘Naturalism’ and ‘Skepticism’ in Hume’s Treatise of Human Na- ture.” Philosophy Compass 3(4), pp. 721–33.

Greig, J.Y.T. 1932/2011. The Letters of David Hume. Oxford: Oxford University Press.

167 Guerrini, Anita. 1985. “James Keill, George Cheyne, and Newtonian Physiology, 1690–1740.” Journal of the History of Biology 18(2), pp. 247–66.

Gupta, Anil. 2019. Conscious Experience: A Logical Inquiry. Cambridge, MA: Harvard University Press.

Guyer, Paul. 1987. “The Failure of the B Deduction.” The Southern Journal of Philoso- phy 25(S1), pp. 67–84.

———. 2003. “Kant on Common Sense and Scepticism.” Kantian Review 7, pp. 1–37.

Haddock, Adrian and Fiona Macpherson, eds. 2008. Disjunctivism: Perception, Action, Knowledge. Oxford: Oxford University Press.

Hamann, Johann Georg. 1955. Briefwechsel, eds. W. Ziesemer and A. Henkel. Wiesbaden: Insel-Verlag.

Harris, James A. 2015. Hume: An Intellectual Biography. Cambridge: Cambridge University Press.

Hatﬁeld, Gary. 2001. “The Prolegomena and the Critiques of Pure Reason.” In Kant Und Die Berliner Aufkl¨arung:Akten des IX Internationalen Kant-Kongresses, eds. Volker Gerhardt, Rolf-Peter Horstmann, and Ralph Schumacher. Walter de Gruyter, pp. 185–208.

———. 2003. “What Were Kant’s Aims in the Deduction?” Philosophical Topics 31(1/2), pp. 165–98.

Hawthorne, John. 2013. “The Case for Closure.” In Steup et al.(2013), pp. 26–43.

Henrich, Dieter. 1969. “The Proof-Structure of Kant’s Transcendental Deduction.” Review of Metaphysics 22(4), pp. 640–59.

———. 1989. “Kant’s Notion of a Deduction.” In F¨orster(1989), pp. 29–46.

Hoppe, Hansgeorg. 1983. Synthesis Bei Kant. Berlin: de Gruyter.

Hume, David. 1739/2007. A Treatise of Human Nature, eds. David Fate Norton and Mary J. Norton. Oxford: Clarendon Press.

———. 1748/2000. An Enquiry Concerning Human Understanding, ed. Tom L. Beauchamp. Oxford: Clarendon Press.

———. 1757/2007. A Dissertation on the Passions: The Natural History of Religion: A Critical Edition, ed. Tom L. Beauchamp. Oxford: Clarendon Press.

———. 1758/1987. Essays, Moral, Political, and Literary, ed. Eugene F. Miller. Indianapolis: Liberty Fund.

168 ———. 1778/1983. The History of England: from the Invasion of Julius Caesar to The Revolution in 1688, ed. William B. Todd. Indianapolis, IN: Liberty Fund.

———. 1779/1947. Dialogues Concerning Natural Religion, ed. Norman Kemp Smith. Ed- inburgh: Thomas Nelson & Sons.

Ichikawa, Jonathan, ed. 2017. The Routledge Handbook of Epistemic Contextualism. Milton Park: Routledge.

Ievers, Lisa. 2015. “The Method in Hume’s “Madness”.” Hume Studies 41(1), pp. 3–32.

Jackson, Stanley W. 1969. “Galen — On Mental Disorders.” Journal of the History of the Behavioral Sciences 5(4), pp. 365–84.

———. 1978. “Melancholia and the Waning of the Humoral Theory.” Journal of the History of Medicine and Allied Sciences 33(3), pp. 367–76.

———. 1990. Melancholia and Depression: From Hippocratic Times to Modern Times. New Haven, CT: Yale University Press.

Jenkins, Joyce, Jennifer Whiting, and Christopher Williams, eds. 2005. Persons and Pas- sions: Essays in Honor of Annette Baier. Notre Dame, IN: University of Notre Dame Press.

Johnson, Samuel. 1756. A Dictionary of the English Language: in which words are deduced from their originals, and illustrated in their diﬀerent signiﬁcations by examples from the best writers. . . . 2nd ed. London: W. Strahan.

Jones, W.H.S. 1931. Hippocrates. Loeb Classical Library, Cambridge, MA: Harvard Univer- sity Press.

Jonson, Ben. 1600. Every Man Out of His Humour. London: William Holme.

Jouanna, Jacques. 2012. Greek Medicine from Hippocrates to Galen: Selected Papers, ed. John Scarborough. Studies in Ancient Medicine, Leiden, Netherlands: Brill.

Kail, Peter. 2008. “On Hume’s Appropriation of Malebranche: Causation and Self.” Euro- pean Journal of Philosophy 16(1), pp. 55–80.

Kant, Immanuel. 1764. “Essay on the maladies of the head.” In Kant(2007), pp. 63–77.

———. 1781–7/1929. Critique of Pure Reason, ed. Norman Kemp Smith. New York: Pal- grave Macmillon.

———. 1783/2004. Prolegomena to Any Future Metaphysics: That Will Be Able to Come Forward as Science, ed. Gary Hatﬁeld. Cambridge: Cambridge University Press.

———. 1788. Critique of Practical Reason. In Kant(1996). pp. 133–271.

169 ———. 1790. “On a discovery whereby any new critique of pure reason is to be made superﬂuous by an older one.” In Kant(2002), pp. 271–336.

———. 1790/2000. Critique of the Power of Judgment, eds. Paul Guyer and Eric Matthews. Cambridge: Cambridge University Press.

———. 1793/1804. “What real progress has metaphysics made in Germany since the time of Leibniz and Wolﬀ?” In Kant(2002), pp. 337–424.

———. 1796. “Proclamation of the imminent conclusion of a treaty of perpetual peace in philosophy.” In Kant(2002), pp. 451–60.

———. 1798. Anthropology from a Pragmatic Point of View. In Kant(2007). pp. 227–429.

———. 1992. Lectures on Logic, ed. J. Michael Young. Cambridge: Cambridge University Press.

———. 1996. Practical Philosophy, ed. Mary Gregor. Cambridge: Cambridge University Press.

———. 1997. Lectures on Metaphysics, eds. Karl Ameriks and Steve Naragon. Cambridge: Cambridge University Press.

———. 1999. Correspondence, ed. Arnulf Zweig. Cambridge: Cambridge University Press.

———. 2002. Theoretical Philosophy After 1781, eds. Henry Allison, Peter Heath, Gary Hatﬁeld, and Michael Friedman. Cambridge: Cambridge University Press.

———. 2007. Anthropology, History, and Education, eds. Robert Louden and G¨unter Z¨oller. Cambridge: Cambridge University Press.

Kemp Smith, Norman. 1941. The Philosophy of David Hume. New York: Palgrave Macmil- lian.

Klibansky, Raymond, Erwin Panofsky, and Fritz Saxl. 1964/2019. Saturn and Melancholy: Studies in the History of Natural Philosophy Religion and Art. Montreal: McGill-Queen’s University Press.

Kopec, Matthew and Michael G. Titelbaum. 2016. “The Uniqueness Thesis.” Philosophy Compass 11(4), pp. 189–200.

Korsgaard, Christine M. 2009. Self-Constitution: Agency, Identity, and Integrity. Oxford: Oxford University Press.

Kuehn, Manfred. 1983. “Kant’s Conception of “Hume’s Problem”.” Journal of the History of Philosophy 21(2), pp. 175–93.

Kvanvig, Jonathan L. 2006. “Closure Principles.” Philosophy Compass 1(3), pp. 256–67.

170 Lear, Jonathan. 1984. “The Disappearing ‘We’.” Proceedings of the Aristotelian Society, Supplementary Volumes 58, pp. 219–58.

Lewis, David. 1996. “Elusive Knowledge.” Australasian Journal of Philosophy 74, pp. 549– 467.

Livingston, Donald. 1998. Philosophical Melancholy and Delirium: Hume’s Pathology of Philosophy. Chicago, IL: University of Chicago Press.

Macarthur, David. 2003. “McDowell, Scepticism, and the ‘Veil of Perception’.” Australasian Journal of Philosophy 81(2), pp. 175–90.

Malcolm, Norman. 1952. “Moore and Ordinary Language.” In Schilpp(1952), pp. 343–68.

———. 1963a. “George Edward Moore.” In Malcolm(1963b), pp. 163–85.

———. 1963b. Knowledge and Certainty. Englewood Cliﬀs, NJ: Prentice-Hall, Inc.

Mandeville, Bernard de. 1715. A treatise of the hypochondriack and hysterick passion. . . . London: Dryden Leab.

Matherne, Samantha. 2015. “Images and Kant’s Theory of Perception.” Ergo: An Open Access Journal of Philosophy 2.

McCain, Kevin and Ted Poston, eds. 2018. The Mystery of Skepticism. Leiden, Netherlands: Brill.

McCann, Edwin. 1985. “Skepticism and Kant’s B Deduction.” History of Philosophy Quar- terly 2(1), pp. 71–89.

McDowell, John. 1982/1998. “Criteria, Defeasibility, and Knowledge.” In McDowell(1998), pp. 369–94.

———. 1994. Mind and World. Cambridge, MA: Harvard University Press.

———. 1995. “Knowledge and the Internal.” Philosophy and Phenomenological Re- search 55(4), pp. 877–93.

———. 1998. Meaning, Knowledge, & Reality. Cambridge, MA: Harvard University Press.

———. 2002. “Knowledge and the Internal Revisited.” Philosophy and Phenomenological Research 64(1), pp. 97–105.

———. 2008/2009. “The Disjunctive Conception of Experience as Material for a Transcen- dental Argument.” In McDowell(2009), pp. 225–40.

———. 2009. The Engaged Intellect: Philosophical Essays. Cambridge, MA: Harvard Uni- versity Press.

171 ———. 2011. “Tyler Burge on Disjunctivism.” Philosophical Explorations 13(3), pp. 243–55.

———. 2013. “Tyler Burge on Disjunctivism (II).” Philosophical Explorations 16(3), pp. 259–79.

———. 2018. “Perceptual Experience and Empirical Rationality.” Analytic Philoso- phy 59(1), pp. 89–98.

McGinn, Marie. 1989. Sense and Certainty: A Dissolution of Scepticism. Oxford: Basil Blackwell.

Meeker, Kevin. 1998. “Hume: Radical Sceptic or Naturalized Epistemologist?” Hume Studies 24(1), pp. 31–52.

———. 2006. “Was Hume a Proper Functionalist?” Philosophy and Phenomenological Research 72(1), pp. 120–36.

Millar, Alan. 2008. “Disjunctivism and Skepticism.” In Greco(2008), pp. 581–601.

Millican, Peter, ed. 2002. Reading Hume on Human Understanding: Essays on the First Enquiry. Oxford: Oxford University Press.

Milton, John. 1674. Paradise lost, a poem in twelve books. 2nd ed. Glasgow: S. Simmons.

Miracchi, Lisa. 2017. “Perspectival Externalism Is the Antidote to Radical Skepticism.” Episteme 14(3), pp. 363–79.

Mizrahi, Moti. 2016. “An Argument for External World Skepticism From the Appear- ance/Reality Distinction.” International Journal for the Study of Skepticism 6(4), pp. 368–83.

Molière.1666/2017. L’Atrabilaire Amoureux: Variation sur Le Misanthrope de Molière, ed. Jacques Kraemer. Paris: Librairie éditionstituli.

———. 1673/2013. Le Malade imaginaire, ed. Georges Couton. Paris: Gallimard.

Moore, Cecil. 1953. Backgrounds of English Literature, 1700–1760. Minneapolis, MI: Uni- versity of Minnesota Press.

Moore, G.E. 1939/1993. “Proof of the External World.” In Moore(1993), pp. 106–33.

———. 1941/1993. “Certainty.” In Moore(1993), pp. 171–96.

———. 1993. G.E. Moore: Selected Writings, ed. Thomas Baldwin. Milton Park: Routledge.

Mossner, Ernest. 1944. “Hume’s Epistle to Dr. Arbuthnot, 1734: The Biographical Signiﬁ- cance.” Huntington Library Quarterly 7, pp. 135–52.

172 ———. 1948. “Hume’s Early Memoranda, 1729–1740: The Complete Text Edited with Foreword.” Journal of the History of Ideas 9(4), pp. 492–518.

Nagel, Thomas. 1978. The Possibility of Altruism. Princeton, NJ: Princeton University Press.

Nietzsche, Friedrich. 1974. The Gay Science: With a Prelude in Rhymes and an Appendix of Songs. New York: Random House.

Norton, David Fate, ed. 1993. The Cambridge Companion to Hume. Cambridge: Cambridge University Press.

Norton, David Fate and Mary J. Norton. 1996. The David Hume Library. Edinburgh: Edinburgh Bibliographical Society & National Library of Scotland.

Nozick, Robert. 1981. Philosophical Explanations. Cambridge, MA: Harvard University Press.

Oxford University. 1971. The Complete Edition of the Oxford English Dictionary. Oxford: Oxford University Press.

Peacocke, Christopher. 2004. The Realm of Reason. Oxford: Oxford University Press.

Porter, Roy. 1987. Mind-forg’d Manacles: A History of Madness in England from the Restoration to the Regency. Cambridge, MA: Harvard University Press.

———. 1999. The Greatest Beneﬁt to Mankind: A Medical History of Humanity. The Norton History of Science, New York: W. W. Norton & Co.

———. 2004. Flesh in the Age of Reason. New York: W.W. Norton & Co.

Pritchard, Duncan. 2008. “McDowellian Neo-Mooreanism.” In Haddock and Macpherson (2008), pp. 283–307.

———. 2009. “Wright Contra McDowell on Perceptual Knowledge and Scepticism.” Syn- these 171(3), pp. 467–79.

———. 2015. Epistemic Angst: Radical Skepticism and the Groundlessness of Our Believing. Princeton, NJ: Princeton University Press.

Pryor, James. 2000. “The Skeptic and the Dogmatist.” Noˆus 34, pp. 517–49.

———. 2004. “What’s Wrong with Moore’s Argument?” Philosophical Issues 14(1), pp. 349–78.

Putnam, Hilary. 1981. Reason, Truth and History. Cambridge: Cambridge University Press.

Qu, Hsueh. 2014. “Hume’s Practically Epistemic Conclusions?” Philosophical Stud- ies 170(3), pp. 501–24.

173 ———. 2020. Hume’s Epistemological Evolution. Oxford: Oxford University Press.

Reed, Phillip A. and Rico Vitz, eds. 2018. Hume’s Moral Philosophy and Contemporary Psychology. New York: Routledge.

Reynolds, Steven L. 2000. “The Argument From Illusion.” Noˆus 34(4), pp. 604–21.

Rinard, Susanna. 2018. “Reasoning One’s Way Out of Skepticism.” In McCain and Poston (2018), pp. 240–64.

Robin, Percy. 1911. The Old Physiology in English Literature. London: J.M. Dent & Sons Ltd.

Rorty, Richard. 1970. “Strawson’s Objectivity Argument.” The Review of Metaphysics 24(2), pp. 207–44.

———. 1979. “Transcendental Arguments, Self-Reference, and Pramatism.” In Bieri et al. (1979), pp. 77–103.

———. 1980. “Pragmatism, Relativism, and Irrationalism.” Proceedings and Addresses of the American Philosophical Association 53(6), pp. 717–38.

Schafer, Karl. 2014. “Curious Virtues in Hume’s Epistemology.” Philosophers’ Imprint 14(1), pp. 1–20.

———. forthcoming. “The Beach of Skepticism: Kant and Hume on the Practice of Phi- losophy and the Proper Bounds of Skepticism.” In The Cambridge Companion to the Prolegomena, ed. Peter Thielke. Cambridge: Cambridge University Press.

Schilpp, Paul A. 1952. The Philosophy of G.E. Moore. New York: Tudor Publishing Co.

Schmid, Jelscha. ms. “The Experiment of Reason: Kant’s Propaedeutic Method.” Unpub- lished manuscript.

Schmitt, Frederick F. 1992. Knowledge and Belief. London: Routledge.

———. 2014. Hume’s Epistemology in the Treatise: A Veritistic Interpretation. Oxford: Oxford University Press.

Schubbach, Arno. 2017. “Kants Konzeption der Geometrischen Darstellung.” Kant- Studien 108(1), pp. 19–54.

Schultheis, Ginger. 2018. “Living on the Edge: Against Epistemic Permissivism.” Mind 127(507), pp. 863–79.

Schulting, Dennis and Jacco Verburgt, eds. 2010. Kant’s Idealism: New Interpretations of a Controversial Doctrine. Dordrecht: Springer.

174 Shakespeare, William. 1609/1905. Shakespeare’s Sonnets, ed. William Rolfe. New York: American Book Company.

Sommerlatte, Curtis. 2016. “Empirical Cognition in the Transcendental Deduction: Kant’s Starting Point and His Humean Problem.” Kantian Review 21(3), pp. 437–63.

Sosa, Ernest. 2007. A Virtue Epistemology: Apt Belief and Reﬂective Knowledge, vol. 1. Oxford: Clarendon Press.

Stern, Robert, ed. 1999. Transcendental Arguments: Problems and Prospects. Oxford: Clarendon Press.

Stern, Robert. 2000. Transcendental Arguments and Scepticism: Answering the Question of Justiﬁcation. Oxford: Oxford University Press.

———. 2006. “Metaphysical Dogmatism, Humean Scepticism, Kantian Criticism.” Kantian Review 11, pp. 102–16.

———. 2008. “Kant’s Response to Skepticism.” In Greco(2008), pp. 265–85.

Steup, Matthias, John Turri, and Ernest Sosa, eds. 2013. Contemporary Debates in Episte- mology. Contemporary Debates in Philosophy, Oxford: Wiley.

Stewart, M.A., ed. 1991. Studies in the Philosophy of the Scottish Enlightenment. Oxford: Clarendon Press.

Strawson, Galen. 2002. “David Hume: Objects and Powers.” In Millican(2002), pp. 231–57.

Strawson, P.F. 1959. Individuals: An Essay in Descriptive Metaphysics. London: Routledge.

———. 1966/2002. The Bounds of Sense. New York: Taylor & Francis.

———. 1985. Skepticism and Naturalism: Some Varieties. New York: Columbia University Press.

Stroud, Barry. 1968/2000. “Transcendental Arguments.” In Stroud(2000), pp. 9–25.

———. 1977. Hume. New York: Routledge.

———. 1984. The Signiﬁcance of Philosophical Scepticism. Oxford: Oxford University Press.

———. 1991/2011. “Hume’s Scepticism: Natural Instincts and Philosophical Reﬂection.” In Stroud(2011), pp. 144–66.

———. 1994/2000. “Kantian Argument, Conceptual Capacities, and Invulnerability.” In Stroud(2000), pp. 155–76.

———. 1999. “The Goal of Transcendental Arguments.” In Stern(1999), pp. 155–72.

175 ———. 2000. Understanding Human Knowledge: Philosophical Essays. Oxford: Oxford University Press.

———. 2006. “The Constraints of Hume’s Naturalism.” Synthese 152(3), pp. 339–51.

———. 2011. Philosophers Past and Present: Selected Essays. London: Clarendon Press.

Thielke, Peter. 2003. “Hume, Kant and the Sea of Illusion.” Hume Studies 29(1), pp. 63–88.

———. 2015. “Turnabout Is Fair Play: A New Humean Response in the Old Debate with Kant.” Hume Studies 41(2), pp. 263–88.

Th¨ole,Bernhard. 1991. Kant und das Problem der Gesetzm¨aßigkeitder Natur. Quellen und Studien zur Philosophie, Berlin: De Gruyter.

Tolley, Clinton. 2020. “Kant on the Place of Cognition in the Progression of Our Represen- tations.” Synthese 197(8), pp. 3215–44.

Traiger, Saul, ed. 2008. The Blackwell Guide to Hume’s Treatise. New York: Wiley.

Walker, Ralph C.S. 1999. “Induction and Transcendental Arguments.” In Stern(1999), pp. 13–29.

Watkins, Eric. 2004. “Kant’s Model of Causality: Causal Powers, Laws, and Kant’s Reply to Hume.” Journal of the History of Philosophy 42(4), pp. 449–88.

Watkins, Eric and Marcus Willaschek. 2020. “Kant on Cognition and Knowledge.” Syn- these 197(8), pp. 3195–3213.

Watkins, Margaret. 2018. “Beyond the ‘Disease of the Learned’: Hume on Passional Disor- ders.” In Reed and Vitz(2018), pp. 9–39.

Waxman, Wayne. 2003. Hume’s Theory of Consciousness. Cambridge: Cambridge University Press.

White, Roger. 2005. “Epistemic Permissiveness.” Philosophical Perspectives 19(1), pp. 445– 59.

Wicker, Ricardo. 2016. Hume’s Solution to Sceptical Melancholy. MA thesis, Universitédu Quebec àMontréal.

Wilkerson, T.E. 1971. “Time, Cause and Object: Kant’s Second Analogy of Experience.” Kant-Studien 62(1-4), pp. 351–66.

Willaschek, Marcus. 2018. Kant on the Sources of Metaphysics: The Dialectic of Pure Reason. Cambridge: Cambridge University Press.

Williams, Michael. 1991. Unnatural Doubts: Epistemological Realism and the Basis of Scep- ticism. Princeton, NJ: Blackwell.

176 ———. 2004. “The Unity of Hume’s Philosophical Project.” Hume Studies 30(2), pp. 265–96.

Williamson, Timothy. 2000. Knowledge and its Limits. Oxford: Oxford University Press.

Willis, Thomas. 1683. Two discourses concerning the soul of brutes: which is that of the vital and sensitive of man. . . . London: Thomas Dring.

Wittgenstein, Ludwig. 1969/1972. On Certainty, eds. G.E.M. Anscombe and G.H. von Wright. New York: Harper & Row.

———. 2005. The Big Typescript: TS 213, eds. C. Grant Luckhardt and Maximillian A.E. Aue. Oxford: Blackwell.

Wolﬀ, Robert Paul. 1981. “Comment on D.P. Dryer.” Philosophical Forum 13(1), p. 27.

Wolterstorﬀ, Nicholas. 1996. John Locke and the Ethics of Belief. Cambridge: Cambridge University Press.

Wright, Crispin. 1985. “Facts and Certainty.” Proceedings of the British Academy 71, pp. 429–72.

———. 1991a. “Scepticism and Dreaming: Imploding the Demon.” Mind 100(1), pp. 87– 116.

———. 2002. “(Anti-)Sceptics Simple and Subtle: G.E. Moore and John McDowell.” Phi- losophy and Phenomenological Research 65(2), pp. 330–48.

———. 2004. “Warrant for Nothing (and Foundations for Free)?” Aristotelian Society Supplementary Volume 78(1), pp. 167–212.

———. 2008. “Comment on John McDowell’s “The Disjunctive Conception of Experience as Material for a Transcendental Argument”.” In Haddock and Macpherson(2008), pp. 390–404.

Wright, Crispin and Annalisa Coliva, eds. 2012. Mind, Meaning, and Knowledge: Themes From the Philosophy of Crispin Wright. Oxford: Oxford University Press.

Wright, John P. 1983. The Sceptical Realism of David Hume. Minneapolis, MI: University of Minnesota Press.

———. 1991b. “Metaphysics and Physiology: Mind, Body, and the Animal Economy in Eighteenth-Century Scotland.” In Stewart(1991), pp. 251–301.

———. 2003. “Dr. George Cheyne, Chevalier Ramsay, and Hume’s Letter to a Physician.” Hume Studies 29(1), pp. 125–41.

177